Comprehensive Summaries of Uppsala Dissertations

from the Faculty of Medicine

Development and Stability of
Antibiotic Resistance
BY
MARIA SJLUND
ACTA UNIVERSITATIS UPSALIENSIS
UPPSALA 2004
Dissertation presented at Uppsala University to be publicly examined in Hrsalen, Dag
Hammarskjlds vg 17, Uppsala, Friday, October 8, 2004 at 13:15 Ior the degree oI Doctor oI
Philosophy (Faculty oI Medicine). The examination will be conducted in English.

Sjlund, M. 2004. Development and Stability oI Antibiotic Resistance. Acta Universitatis

Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of
Meaicine. 63 pp. Uppsala. ISBN 91-554-6026-7
Antibiotic resistance is oI current concern. Bacteria have become increasingly resistant to
commonly used antibiotics and we are Iacing a growing resistance problem. The present
thesis was aimed at studying the impact oI antibiotic treatment on pathogenic bacteria as well
as on the normal human microbiota, with Iocus on resistance development.
Among the Iactors that aIIect the appearance oI acquired antibiotic resistance, the mutation
Irequency and biological cost oI resistance are oI special importance. Our work shows that the
mutation Irequency in clinical isolates oI Helicobacter pylori was generally higher than Ior
other studied bacteria such as Enterobacteriaceae, / oI the isolates displayed a mutation
Irequency higher than Enterobacteriaceae deIective mismatch repair mutants and could be
regarded as mutator strains.
In H. pylori, clarithromycin resistance conIers a biological cost, as measured by decreased
competitive ability oI the resistant mutants in mice. In clinical isolates, this cost could be
reduced, consistent with compensatory evolution stabilizing the presence oI the resistant
phenotype in the population. Thus, compensation is a clinically relevant phenomenon that can
occur in vivo.
Furthermore, our results show that clinical use oI antibiotics selects Ior stable resistance in
the human microbiota. This is important Ior several reasons. First, many commensals
occasionally can cause severe disease, even though they are part oI the normal microbiota.
ThereIore, stably resistant populations increase the risk oI unsuccessIul treatment oI such
inIections. Second, resistance in the normal microbiota might contribute to increased
resistance development among pathogens by interspecies transIer oI resistant determinants.
Keyworas. antibiotic resistance, selection, mutation Irequency, biological cost oI resistance,
compensatory evolution, Helicobacter pylori, normal microbiota
Maria Sfluna, Department of Meaical Sciences, Clinical Bacteriology, Box 552, Uppsala
University, SE-75122 Uppsala, Sweaen
Maria Sjlund 2004
ISSN 0282-7476
ISBN 91-554-6026-7
urn:nbn:se:uu:diva-4523 (http://urn.kb.se/resolve?urnurn:nbn:se:uu:diva-4523)
List of Papers
This thesis is based on the following papers, which are referred to in the text
by their roman numerals:
I. Bjrkholm B, Sjlund M, Falk P, Berg O, Engstrand L and Anders-
son DI, Mutation frequency and biological cost of resistance in Helicobacter
pylori. Proc Natl Acad Sci U S A 2001 Dec 4;98(25):14607-12
II. Gustafsson I, Sjlund M, Torell E, Johannesson M, Engstrand L,
Cars O and Andersson DI, Bacteria with increased mutation frequency and
antibiotic resistance are enriched in the commensal flora of patients with
high antibiotic usage, Journal of Antimicrobial Chemotherapy (2003) 52,
645650.
III. Sjlund M, Wreiber K, Andersson DI, Blaser MJ and Engstrand L,
Long-term persistence of resistant Enterococcus species after antibiotic
treatment to eliminate Helicobacter pylori. Annals of Internal Medicine
2003;139:483-487
IV. Sjlund M, Tano E, Blaser MJ, Andersson DI and Engstrand L,
Fitness of antibiotic resistant Staphylococcus epidermidis. In manuscript.
Reprints were made with permission from the publisher.
Contents
The Antibiotic Resistance Problem.................................................................9
THE INTRODUCTION OF ANTIBIOTICS...........................................10
Mechanisms of Action of Antibiotics..................................................11
ANTIBIOTIC RESISTANCE..................................................................13
Mechanisms of Antibiotic Resistance..................................................13
Acquired Antibiotic Resistance ...........................................................14
SELECTION OF ANTIBIOTIC RESISTANT BACTERIA...................16
The Volume of Drug Use ....................................................................16
The Rate of Formation of Resistant Mutants.......................................17
Transmission of Resistant Bacteria .....................................................19
STABILITY OF ANTIBIOTIC RESISTANT BACTERIA....................20
The Biological Cost of Antibiotic Resistance......................................21
Compensatory Evolution .....................................................................22
FURTHER FACTORS INFLUENCING THE STABILITY OF
ANTIBIOTIC RESISTANCE..................................................................24
Genetic Linkage of Resistance Genes .................................................25
No-cost Associated Resistance Mutations...........................................25
THE ROLE OF THE HUMAN NORMAL MICROBIOTA IN
ANTIBIOTIC RESISTANCE DEVELOPMENT ...................................26
The Human Microbiota........................................................................26
GENERAL AIM...........................................................................................29
SPECIFIC AIMS ..........................................................................................30
MATERIALS AND METHODS..................................................................31
Bacterial Species Studied in the Present Thesis.......................................31
Helicobacter pylori ..............................................................................31
Enterococcus spp. ................................................................................31
Staphylococcus epidermidis.................................................................33
-streptococci ......................................................................................34
Escherichia coli ...................................................................................34
Bacterial Strains (I-IV).............................................................................35
DNA Preparation and Polymerase Chain Reaction (I-IV) .......................36
Determination of Antibiotic Susceptibilities (II)......................................37
Minimal Inhibitory Concentration Determinations (MIC) (I-IV) ............37
Mutation Frequency Determination (I, II)................................................37
Sequencing (II).........................................................................................38
Statistical Analysis (II).............................................................................39
Mathematical Modelling (I) .....................................................................39
Appendix 1 ..........................................................................................39
Appendix 2 ..........................................................................................40
In vivo Competition Model (I) .................................................................41
Pulsed-Field Gel Electrophoresis (PFGE) (IV)........................................41
Competition Assay on Human Skin (IV) .................................................42
RESULTS AND DISCUSSION...................................................................43
Mutator strains are common among clinical isolates of Helicobacter
pylori (I). ..................................................................................................43
Antibiotic pressure is a selector for elevated mutation frequencies (II)...45
Clarithromycin resistance confers a biological cost in Helicobacter pylori
but can be reduced by compensatory evolution in vivo (I) ......................48
Antibiotic treatment selects for stable resistant strains in the human
normal microbiota (II, III, IV)..................................................................49
CLINICAL IMPLICATIONS.......................................................................52
ACKNOWLEDGEMENTS..........................................................................53
REFERENCES .............................................................................................56
Abbreviations
AP-PCR Arbitrary primed PCR
ATC-system The Anatomical Therapeutic Chemical
system
bp Base pairs
CFU Colony forming units
CI Competition index
ClaS Clarithromycin-susceptible
ClaR Clarithromycin-resistant
CoNS Coagulase-negative staphylococci
kb Kilo base pairs
DDD Defined daily dose
DNA Deoxyribonucleic acid
FCS Fetal calf serum
LB Luria-Bertani
MIC Minimal inhibitory concentration
MMR Mismatch repair
MRSA Methicillin-resistant Staphylococcus
aureus
PCR Polymerase chain reaction
PFGE Pulsed-field gel electrophoresis
PBP Penicillin binding protein
PBS Phosphate buffered saline
RifR Rifampicin-resistant
RNA Ribonucleic acid
SmR Streptomycin-resistant
VRE Vancomycin-resistant enterococci
9
The Antibiotic Resistance Problem
Bacteria have become increasingly resistant to commonly used antibiotics,
and we are facing a growing resistance problem (Finch 1998). Before the
introduction of antibiotics, the hospital pathogens of major concern were
Staphylococcus aureus and Group A-streptococci (Williams 2001). Today,
however, the physicians are facing infections caused by resistant Gram-
negatives, multi-resistant S. aureus, coagulase-negative staphylococci,
pneumococci and enterococci (Baquero 1997; Williams 2001; Clark, Her-
shberger et al. 2003). Thus, infections that were readily cured by antibiotics
in the past may today be difficult or impossible to treat. This threat has
brought forward the concept of a post-antimicrobial era, in which some
infections would no longer be susceptible to antibiotic therapy (Cohen
1992). Since antibiotics not only are important for treating specific infec-
tions, but also have a profound impact on many other aspects of medicine,
such as oncology and transplantation surgery, the post-antimicrobial era
represents a worst-case scenario similar to the pre-antibiotic era, when mor-
bidity and mortality associated with infectious diseases were high.
Most reports suggest the emergence of resistant bacteria is the price we
have to pay for an inappropriate use of antibiotics during the last decades,
and that the overuse and misuse of antibiotics constitute the major force be-
hind the appearance and spread of resistance (Levy 2001). However, devel-
opment of antibiotic resistance is a multifaceted problem which is dependent
on several factors (Barbosa and Levy 2000; Low 2001), including:
(i) The volume of drug use
(ii) The rate of formation of resistant mutants
(iii) The biological cost of resistance and to what extent compen-
satory evolution may act to reduce such a cost
In the present thesis, each one of these factors will be addressed and their
relevance in the development and spread of antibiotic resistance discussed.
10
THE INTRODUCTION OF ANTIBIOTICS
With the introduction of the sulphonamides in the 1930s, followed by peni-
cillin in the 1940s, the antimicrobial era had begun. Considered one of the
most important events in medical history, the discovery of antibiotics revolu-
tionized the field of infectious diseases by giving the physicians the ability to
prevent, cure and reduce transmission of certain diseases. Consequently, a
significant reduction in morbidity and mortality associated with infectious
disease had been achieved (Cohen 2000).
The success of penicillin in the 1940s led researchers to intensify the
search for new antibiotics that could treat other bacterial diseases. Therefore,
during the 1940s to the beginning of the 1970s, the development and pro-
duction of antimicrobial compounds was very successful, resulting in several
new classes of antibiotics. However, during the 1970s the production of
antibiotics declined, and it took almost 30 years before a new class of antibi-
otics the oxazolidinones (Norrby 2001; Moellering 2003) was introduced
on the market. Instead, during this period of time, the new antibiotics intro-
duced primarily consisted of chemical modifications of already known com-
pounds.
Figure 1. The introduction of different antibiotics. Adapted from Norrby and Cars,
Antibiotika- och Kemoterapi, 2003 (Norrby 2003).
Oxazolidinones
Sulphonamides
Penicillin
Trimethoprim
Glycopeptides
Macrolides
Chloramphenicol
Tetracyclines
Quinolones
Lincosamides
Streptogramins
1930 1940 1970 1960 1950 2000
Aminoglycosides
11
Mechanisms of Action of Antibiotics
Antibiotics can be classified in several ways. One common method of classi-
fication is by their mechanism of action against the infecting bacteria. Some
antibiotics act by interfering with the synthesis of proteins and nucleic acids
in the bacteria, while others attack the cell wall or disrupt the cell membrane
(Green 2002). A clinically important group of antibiotics interferes with the
synthesis of the peptidoglycan, the most important component of the cell
wall. This group of antibiotics is called the -lactams and can further be
divided into the penicillins, cephalosporins, monobactams and carbapenems
(Green 2002). Another large group of antibiotics inhibits the synthesis of
various intracellular molecules, such as DNA, RNA, ribosomes and proteins.
Examples of such antibiotics are rifampicin, which inhibits the RNA poly-
merase, and the quinolones (Blondeau 2004), which inhibit the enzymes
responsible for coiling and uncoiling the DNA molecule, a process necessary
for DNA replication and transcription. There are also other mechanisms;
macrolides interfere with the 50S subunit of the ribosome, whereas tetracy-
clines affect the 30S ribosomal subunit, both inhibiting protein synthesis.
Table.1 Mechanisms of action of different antibiotic classes (Schmid 2001).
Antibiotic class Molecular target Examples

-lactams Cell wall; Penicillin bind-
ing proteins
Penicillins (benzylpenicil-
lin, ampicillin, amoxycil-
lin)
Cephalosporins, (cefo-
taxime, ceftazidime)
Carbapenems (imipenem,
meropenem)
Monobactams (aztreonam)
Aminoglycosides 30S, 50S ribosomal sub-
units
Gentamicin, tobramycin,
amikacin, streptomycin
Trimethoprim, sulphon-
amides
Folate synthesis Trimethoprim, sulfadiaz-
ine
Quinolones Gyrase, Topoisomerase IV Nalidixic acid, Fluoroqui-
nolones (ciprofloxacin,
norfloxacin, moxifloxacin)
Macrolides 50S ribosomal subunit Erythromycin, clarithro-
mycin, azithromycin
12
Table.1 continued.
Antibiotic class Molecular target Examples

Lincosamides 50S ribosomal subunit Clindamycin
Streptogramins 50S ribosomal subunit Quinopristin/Dalfopristin
Tetracycline 30S ribosomal subunit Tetracycline, doxycycline
Glycopeptides Cell wall peptidoglycan Vancomycin, teicoplanin
Chloramphenicol 50S ribosomal subunit Chloramphenicol
Rifamycins RNA polymerase Rifampicin
Polymyxin Cell membrane Polymyxin B, colistin
Oxazolidinone 50S ribosomal subunit Linezolid
Antibiotics may also be classified as bactericidal or bacteriostatic
(Stratton 2003). In general, antibiotics attacking the cell wall belong to the
group of bactericidal drugs, since a defective cell wall eventually will cause
the bacteria to lyse and die. Among the bacteriostatic antibiotics, macrolides,
lincosamides and chloramphenicol can be mentioned. With bacteriostatic
drugs, the host immune system plays an important role, helping to clear the
infection once bacterial growth has subsided. For that reason, bactericidal
drugs should be considered in patients that are immuno-compromised, as
well as in patients with serious infections such as endocarditis and meningi-
tis, in which cases a fast reduction of bacteria is warranted (Pankey and Sa-
bath 2004).
13
ANTIBIOTIC RESISTANCE
For many years antibiotics seemed to be winning the war against infectious
disease. However, despite the successful development of several different
antibiotic classes, the introduction of a new drug was almost always fol-
lowed by resistance. Shortly after the introduction of penicillin, resistance
was detected in Staphylococcus aureus, and by 1970 most S. aureus isolates
were penicillin-resistant (Chambers 2001). In a similar manner, clinicians
soon witnessed clinical failure of other antibiotics due to bacterial resistance
development. For every decade to follow, bacteria resistant not only to single
but multiple antibiotics have become more and more widespread (Tenover
and Hughes 1996). Today, we are facing a problem of multi-resistant Salmo-
nella, Shigella, Campylobacter, Pseudomonas aeruginosa and Mycobacte-
rium tuberculosis, penicillin-resistant pneumococci, vancomycin-resistant
enterococci and methicillin- and vancomycin-resistant Staphylococcus
aureus (Hand 2000; Jones 2001; Lieberman 2003). At the same time fewer
antibiotics are being produced, and it is becoming more and more apparent
that a careful and prudent use of antibiotics is necessary in order to curtail
the development of bacterial resistance (Levy 2001).
Mechanisms of Antibiotic Resistance
In basic terms, the increase in prevalence of antibiotic resistant bacteria dur-
ing the last decades can be attributed to evolution and natural selection. All
populations of organisms, including bacteria, will include variants with un-
usual traits, in this case the ability to withstand antibiotics. Consequently,
every time a specific antibiotic is used, the antibiotic resistance trait will be
positively selected, and the bacteria carrying this trait will increase in num-
ber and eventually predominate the population.
The bacterial traits of antibiotic resistance may be due to several different
mechanisms (Normark and Normark 2002), including:
(i) a decreased uptake of the drug
(ii) an increased export of the drug
(iii) inactivation or modification of the drug target
(iv) the introduction of a new drug resistant target
(v) hydrolysis of the antibiotic
(vi) modification of the drug
(vii) prevention of activation of the drug
14
How do the bacteria acquire these traits? First, antibiotic resistance traits
are naturally occurring in the environment and have been so since long be-
fore antibiotics were introduced into human medicine (Davies 1994). One
theory for the presence of antimicrobial resistance genes in the environment
is that they originate from bacteria or fungi that use them as protection from
antibiotics produced by other bacteria (Hawkey 2000). Another theory for
the source of antibiotic resistance determinants is that certain housekeeping
genes, such as sugar kinases and acetyltransferases, may have evolved to
modify antibiotics, as in the case of aminoglycoside resistance (Davies
1994). Second, some bacteria are naturally resistant to certain antibiotics on
account of their genetic composition. For example, Mycoplasma spp. is al-
ways resistant to -lactams, since this species lacks the peptidoglycan in the
cell wall. Bacteria carrying such resistant traits are designated to be intrinsi-
cally resistant, i.e. naturally resistant to an antibiotic without any genetic
alterations. Finally, bacteria may be genetically altered to become resistant, a
process called acquired resistance (Normark and Normark 2002).
Acquired Antibiotic Resistance
Bacteria can acquire resistance by either of two mechanisms;
(i) spontaneous mutation
(ii) horizontal transfer
Mutations may render the bacteria resistant by modifying the drug target
(mutations in ribosomal proteins, penicillin-binding proteins etc.), by chang-
ing the uptake of the drug (mutations in a porin) or by inducing an increased
efflux of the drug (mutations causing overexpression of efflux pumps)
(Hooper 2001; Normark and Normark 2002).
Spontaneous mutation is dependent on the mutation rate and the presence
of proofreading and repair mechanisms (Miller 1996). Some strains display
an extremely high mutation rate and are called mutator strains. These bacte-
ria are usually defective in the mismatch repair system or lack the ability of
proofreading (Miller 1996; Bridges 2001). The role that mutator strains
might play in generating and speeding up antibiotic resistance development
is discussed further in the chapter Rate of formation of resistant mutants.
Horizontal transfer is a mechanism that allows bacteria to share genetic
material and thereby maintain genetic diversity (Maiden 1998). It also con-
stitutes the main mechanism for acquiring antibiotic resistance determinants.
Essentially, three different processes are involved in horizontal gene trans-
fer: conjugation, transduction and transformation (Rice 2000). Since these
mechanisms can occur not only within the same but also within different
species, horizontal transfer constitutes a major force behind the spread of
resistance (Salyers and Amabile-Cuevas 1997; Maiden 1998). The manner
15
by which horizontal transfer renders bacteria resistant is primarily via the
introduction of new antibiotic targets. This is commonly seen as the recruit-
ment of new genes carried on plasmids or transposons. Thus, as opposed to
spontaneous mutation, the resistance determinant is pre-existing in a reser-
voir and is not the direct result of antibiotic selection of mutants from within
an entirely susceptible bacterial population.
A new antibiotic target could also be introduced by transformation of
DNA and a subsequent recombination into the chromosome. An example of
a resistance determinant that has originated from horizontal transfer and
transformation, is the development of mosaic genes of penicillin-binding-
proteins in S. pneumoniae, conferring penicillin resistance (Hakenbeck
1999). Few human pathogens have this ability; most other clinically impor-
tant pathogens become penicillin-resistant due to the acquisition of genes
encoding -lactamases, which inactivate the -lactams.
Table 2. Major bacterial pathogens and resistance patterns (Burman 2001).
Bacterial pathogen Antibiotic resistance
Escherichia coli -lactam resistance due to plasmid-mediated -
lactamases, trimethoprim resistance, quinolone resis-
tance due to mutations in gyrA
Klebsiella -lactam resistance due to plasmid-mediated -
lactamases
Helicobacter pylori Macrolide and metronidazole resistance
Enterococci Glycopeptide resistance, aminoglycoside resistance,
quinolone resistance, penicillin/carbapenem resistance
(mainly E. faecium)
Staphylococcus aureus Methicillin resistance (MRSA) often combined with
multidrug resistance (aminoglycosides, macrolides,
tetracyclines)
Streptococcus pneumoniae Penicillin resistance combined with multidrug resis-
tance (tetracycline, macrolides, chloramphenicol)
Streptococcus pyogenes Macrolide resistance, tetracycline resistance
Haemophilus influenza -lactam resistance mediated by -lactamases or target
modification
Mycobacteria Multidrug resistance (rifampicin, isoniazid and others)
16
SELECTION OF ANTIBIOTIC RESISTANT
BACTERIA
Several factors will have an impact on how rapidly resistance evolves in a
bacterial population. Of special importance is the selective pressure associ-
ated with the volume of drug use and the rate by which bacteria may acquire
and develop resistance.
The Volume of Drug Use
As previously mentioned, the increased antibiotic usage is thought to be one
major force behind resistance development (Austin, Kristinsson et al. 1999).
With few exceptions, resistance has evolved to all antibiotics after a couple
of years of clinical use. The correlation between antibiotic use and increased
resistance is well established and has been reported in several studies
(Arason, Kristinsson et al. 1996; Granizo, Aguilar et al. 2000; Bronzwaer,
Cars et al. 2002). For example, in Iceland, a strong correlation between anti-
biotic use in the community and carriage of penicillin-resistant pneumococci
in children was described (Arason, Kristinsson et al. 1996), and in the Neth-
erlands, an increased prescription of fluoroquinolones for urinary tract infec-
tions was associated with increased resistance to norfloxacin (Goettsch, van
Pelt et al. 2000). The link between the use of antibiotics and the selection of
resistance was further demonstrated by Asensio et al. who showed that the
number of days patients received antibiotics was correlated to the risk of
getting colonized or infected by methicillin-resistant S. aureus (MRSA)
(Asensio, Guerrero et al. 1996).
Because of these observations, a decreased use of antibiotics has become
one leading strategy in order to reduce and limit the further spread of resis-
tance, although clinical evidence supporting the idea that a reduction in anti-
biotic use results in a reduced frequency of resistance is rather weak. Only
two studies have provided support for the reversibility of antibiotic resis-
tance: Seppl et al. showed that a decrease in macrolide consumption in
Finland led to a significant decrease in the number of macrolide resistant
Streptococcus pyogenes, and in Iceland the incidence of penicillin-resistant
pneumococci declined after a reduction of antibiotic use (Seppl, Klaukka
et al. 1997; Austin, Kristinsson et al. 1999). These data have, however, been
challenged by the fact that other factors, such as clonal shifts towards more
susceptibility in the bacterial population, might have caused the apparent
correlation between reduced antibiotic use and decreased frequency of resis-
tant strains (Andersson 2003). Nevertheless, it is generally agreed that the
development of resistance is faster than the rate by which resistant isolates
decline, and the volume of drug use is definitely an important parameter
affecting the emergence of resistance.
17
The consumption of antibiotics varies between different countries (Cars,
Mlstad et al. 2001). For example, in the European Union, the outpatient use
in 1997 varied between 9 and 36.5 DDD/1000 inhabitants /day, and the
countries with the highest and lowest consumption were France and the
Netherlands, respectively. A relationship between antibiotic use and preva-
lence of resistant strains can be seen within the European Union where the
amount of resistant strains follows a south to north gradient, with the south-
ern European countries having a higher prevalence (Cars, Mlstad et al.
2001).
In Europe, antibiotic resistance development has been monitored by the
European Antimicrobial Resistance Surveillance System (EARSS) since
1998, and attempts to reduce the use of antibiotics by minimizing inappro-
priate antibiotic use are being made in several countries (Cars, Mlstad et al.
2001; Ball, Baquero et al. 2002). A corresponding system called NARMS
(National Antibiotic Resistance Monitoring System) is monitoring the emer-
gence and spread of antibiotic resistant strains in the US and is run by the
Centers for Disease Control and Prevention in Atlanta.
In order to be able to compare drug utilization data from different coun-
tries, the data need to be collected and presented in a standardized way. In
1996 the ATC/DDD system was introduced by the WHO and recommended
as an international tool for presenting drug utilization statistics. In the Ana-
tomical Therapeutic Chemical (ATC) system, drugs are divided into differ-
ent groups according to their chemical, pharmacological and therapeutic
properties. Thus, in the ATC system, all drugs will be given a specific code.
In the DDD system, the defined daily dose (DDD) of all ATC-classified
drugs is established. The defined daily dose is defined as the assumed aver-
age dose per day for a drug used for its main indication in adults and the
drug consumption is often expressed as DDDs/1000 inhabitants per day or,
when considering drug utilization in hospitals, DDDs per 100 bed days.
The Rate of Formation of Resistant Mutants
The rate by which spontaneous mutation and horizontal transfer render bac-
teria resistant will also affect the resistance development.
Mutations occur naturally in bacterial populations and can be found at a
frequency of 10
-9
10
-12
per base pair replicated (Bridges 2001). In order to
counteract the accumulation of mutations, bacteria have developed two main
mechanisms that serve to correct mismatching nucleotides and increase the
fidelity. First, the DNA-polymerase exhibits a proofreading activity, which
involves a 3- to 5 exonuclease effect that removes incorrectly paired bases
in the DNA. Second, bacteria possess a post-replicative system called the
mismatch-repair system (MMR), which recognizes the mismatched base,
excises it from the newly synthesized strain and restores the sequence with
the correct base (Miller 1996).
18
Some bacteria display an unusually high mutation rate due to the lack of a
functional mismatch repair system. These bacteria are classified as mutator
strains and may have up to a 1000-fold higher mutation rate than the wild
type (Miller 1996). Such strains have been reported among natural isolates
of several different pathogens: Escherichia coli, Salmonella typhimurium,
Pseudomonas aeruginosa and Helicobacter pylori (see Paper I in the present
thesis) (LeClerc, Li et al. 1996; Oliver, Canton et al. 2000; Bjrkholm, Sj-
lund et al. 2001). Since mutation plays a central role in the evolution of anti-
biotic resistance, the presence of mutators in a bacterial population could
enhance the generation of resistance mutations (Giraud, Matic et al. 2002;
Chopra, O'Neill et al. 2003; Miller, O'Neill et al. 2004). Evidence supporting
this theory has been presented: E. coli mutators have been shown to develop
resistance to rifampicin and ciprofloxacin up to 1000-fold faster than normal
strains (Miller, O'Neill et al. 2002). It has further been shown that an E.coli
mutD mutator carrying the -lactamase TEM-1 easily can mutate to obtain
the extended spectrum -lactamase TEM-52 (Orencia, Yoon et al. 2001).
This requires three successive mutations and indicates that an increased mu-
tation rate could facilitate the emergence of resistance.
Thus, a correlation between a high bacterial mutation rate and faster resis-
tance development has been established in laboratory media and in animal
experiments (Tanabe, Kondo et al. 1999; Giraud, Matic et al. 2002; Schaaff,
Reipert et al. 2002). However, whether resistance development is faster in a
patient infected with a mutator strain is still unclear even though it has been
addressed in a few studies. Oliver et al. showed that P. aeruginosa mutator
strains isolated from cystic fibrosis patients were more resistant than wild
type strains. Additionally, Komp-Lindgren et al. presented evidence that a
high mutation rate in clinical isolates of E. coli strongly correlated with
fluoroquinolone resistance (Oliver, Canton et al. 2000; Komp Lindgren,
Karlsson et al. 2003). As part of the present thesis, we investigated whether
an elevated mutation frequency was correlated to increased resistance devel-
opment among commensal isolates of E. coli, enterococci, coagulase-
negative staphylococci and -streptococci (paper II) (Gustafsson, Sjlund et
al. 2003). As opposed to the findings of Oliver et al., we did not detect a
correlation between an increased mutation frequency and resistance devel-
opment for four different classes of antibiotics (aminoglycosides, -lactams,
macrolides and trimethoprim-sulphmethoxazole). However, when comparing
the mutation frequency of ciprofloxacin-resistant and -susceptible isolates a
correlation was found for resistant E. coli. This is in accordance with the
findings of Komp-Lindgren and indicates that a high mutation rate may play
a significant role in generating resistance caused by point mutations.
Thus, mutator strains are more likely to accumulate mutations, like resis-
tance mutations, but they are also more likely to go extinct because of ac-
quired lethal mutations (Giraud, Radman et al. 2001). However, both theo-
retical and experimental data have shown that, under certain conditions, it
19
can be beneficial to be a mutator (Mao, Lane et al. 1997; Sniegowski, Ger-
rish et al. 1997; Taddei, Matic et al. 1997; Taddei, Radman et al. 1997). Be-
cause of their high mutation rate, mutator strains can more easily adapt to
changes in the environment (Tenaillon, Toupance et al. 1999). Thus, a muta-
tor can reach a higher density in association with an adaptive mutation, given
that the adaptive mutation is at a selective advantage (Arjan and de Visser
2002). Such hitchhiking of mutators could be one potential explanation to
why mutators have been found so frequently among isolates of P. aerugi-
nosa, H. pylori, E. coli and S. typhimurium.
The rate by which antibiotic resistance mutations are formed is further
dependent on the target size, i.e. how many genes and base substitutions that
can confer resistance (Martinez and Baquero 2000). For instance, in E. coli
seven point mutations in the gene gyrA will result in fluoroquinolone resis-
tance, whereas only three mutations in parC leads to resistance (Hooper
1999; Martinez and Baquero 2000). Consequently, the mutation rate for the
gyrA gene will be higher. However, the overall mutation rate of fluoroqui-
nolone resistance in E. coli is determined by the sum of all possible ways of
achieving resistance.
Moreover, external conditions, such as exposure to toxic compounds,
might affect the mutation rate to resistance. For instance, certain types of
antibiotics (fluoroquinolones and aminoglycosides) have been shown to ex-
hibit a mutagenic activity and may lead to increased resistance development
(Ysern, Clerch et al. 1990; Ren, Rahman et al. 1999).
Horizontal gene transfer can mediate resistance by the exchange of plas-
mids, transposons or chromosomal DNA. Some bacteria are naturally trans-
formable, i.e. can take up naked DNA very easily. Such bacteria can acquire
resistance either by recruiting a new gene encoding an alternative, antibiotic-
resistant target molecule, or by intragenic recombination between related
genes that results in novel alleles that are mosaic genes and encode resistant
proteins. The formation of resistant mosaic genes are of special importance
for S. pneumoniae and Neisseria which have evolved penicillin resistance by
mosaic genes encoding PBPs with low affinity for -lactams (Hakenbeck
1999).
Transmission of Resistant Bacteria
Once resistance to an antibiotic has appeared in a bacterial population, the
next question is whether the resistant bacteria will be able to survive and
spread. The success of the bacteria will partly be determined by bacterial
factors, that is, the bacterias epidemic properties to colonize and be trans-
mitted to new hosts. These bacterial properties are mainly dependent on the
bacterial fitness and will be discussed further in the next chapter; Stability of
resistant bacteria. Moreover, a number of non-bacterial factors may influ-
20
ence the dissemination of bacteria. For example, hospital hygiene and infec-
tion control measurements have played a crucial role in combating resistance
and will continue to have a central role in preventing the spread of resistance
in hospitals, especially in countries where multi-drug resistant bacteria such
as MRSA and VRE are common (Pittet 2003).
STABILITY OF ANTIBIOTIC RESISTANT
BACTERIA
The stability and maintenance of a resistant population is mainly deter-
mined by the fitness and transmission costs of resistance (Andersson and
Levin 1999; Bjrkman and Andersson 2000). More specifically, this is de-
pendent on the relative rates by which resistant and susceptible bacteria
(i) grow and die within and outside hosts
(ii) are transmitted between hosts
(iii) are cleared from infected hosts
In order to predict the stability and persistence of antibiotic resistant bac-
teria, all these parameters would be very valuable to know. However, deter-
mining these parameters is in practice very difficult. Instead, we choose to
measure the growth rate of the bacteria, in vitro and in vivo, which will re-
flect the rates of exponential growth of the bacteria, their resource utilization
efficiencies and their mortality in the presence and absence of a host. By
studying the growth rates of resistant bacteria and their competitive perform-
ance against susceptible bacteria, we can get a measure of the relative fitness
of the resistant bacteria and thereby predict the stability and persistence of
the resistant population (Andersson and Levin 1999).
There is no doubt that bacteria benefit from possessing an antibiotic resis-
tance gene as long as the antibiotic is present. But what happens when the
selective pressure of antibiotics is removed? If carrying a resistance gene
confers a reduction in fitness of the bacteria in the absence of an antibiotic
pressure, then one strategy to combat resistance development would be to
stop using the antibiotic until the resistant bacteria declined to a low fre-
quency. A number of studies have indeed presented evidence that resistant
bacteria generally are less fit than their susceptible counterparts. This fitness
reduction is commonly described as the biological cost of resistance and is
considered as a key parameter in determining the rate of ascent and dissemi-
nation of antibiotic resistance.
21
The Biological Cost of Antibiotic Resistance
Data from a number of laboratory studies indicate that most resistance-
conferring mutations are associated with a fitness cost in bacteria (Bjrkman,
Samuelsson et al. 1999; Reynolds 2000; Bjrkholm, Sjlund et al. 2001;
Nagaev, Bjrkman et al. 2001). This cost can be explained by the fact that
the majority of resistance mutations occur in genes that have essential func-
tions in the cell and may alter or impair the function of the target. For exam-
ple, rpoB mutations in E. coli affect the rate of transcription (Reynolds 2000)
and in a similar manner, mutations in the fusA gene in S. typhimurium
(Bjrkman, Nagaev et al. 2000) and S. aureus (Nagaev, Bjrkman et al.
2001) or rpsL mutations in E. coli (Levin, Perrot et al. 2000) and S. typhimu-
rium (Bjrkman, Nagaev et al. 2000) decrease the rate of translation, and as
a consequence, the growth rate is reduced. Also, accessory elements carrying
resistance genes may confer a cost, which can be related to the replication
and maintenance of the elements themselves (Bjrkman and Andersson
2000).
Essentially, there are three ways of measuring the biological cost of resis-
tance in bacteria:
(i) Retrospectively, by studying the relationship between antibiotic
use and resistance in hosts (Seppl, Klaukka et al. 1997; Aus-
tin, Kristinsson et al. 1999)
(ii) Prospectively, by measuring the rates at which humans become
infected and are cleared of resistant and susceptible bacteria
(iii) Experimentally, by estimating the relative rates of growth, sur-
vival, clearence and transmission of susceptible and resistant
bacteria in vitro and in vivo. (Schrag, Perrot et al. 1997; Bjrk-
man, Hughes et al. 1998; Bjrkman, Samuelsson et al. 1999;
Reynolds 2000; Bjrkholm, Sjlund et al. 2001; Gustafsson,
Cars et al. 2003)
The majority of data indicating that antibiotic resistance is costly for the
bacteria has been obtained experimentally, and is based on the relative rates
of growth and competitive performance between antibiotic-susceptible and
resistant bacteria. Bjrkman et al. showed that mutations in rpsL, rpoB and
gyrA causing streptomycin, rifampicin and nalidixic acid resistance, respec-
tively, all confer a cost in S. typhimurium, in vitro as well as in vivo
(Bjrkman, Hughes et al. 1998). Cost-associated resistance mutations have
further been described for fusA mutations causing fusidic acid resistance in
S. aureus and for isoniazid-resistant M. tuberculosis (Li, Kelley et al. 1998).
The greatest limitation when studying the biological cost of resistance ex-
perimentally is how to interpret negative results, i.e. when no cost seems to
22
be associated with the resistance. If no cost can be detected during several
different competition experiments, performed in vitro and in vivo, it is likely
that the cost indeed is low or non-existent. It is, however, of great impor-
tance to perform the experiments both in vitro and in vivo since some muta-
tions may engender a large cost in vitro but no cost in vivo and vice versa
(Nagaev, Bjrkman et al. 2001).
In conclusion, many chromosomal resistance mutations engender a fitness
burden on the bacteria. Accordingly, in the absence of an antibiotic pressure,
resistant bacteria should be less competitive than their susceptible counter-
parts and decline to a low frequency, but that is not the case: bacteria can
overcome the cost of resistance by evolving adaptations that restore the
original fitness.
Compensatory Evolution
To avoid the cost associated with resistance, bacteria may adapt very fast to
their newly acquired trait. The bacteria do this by genetic compensation, i.e.
by introducing mutations that reduce the cost of maintaining the resistance
gene/mutation (Levin, Perrot et al. 2000). This phenomenon has been de-
scribed as compensatory evolution, and has been reported to occur in vitro
(Schrag, Perrot et al. 1997; Bjrkman, Hughes et al. 1998; Bjrkman,
Samuelsson et al. 1999; Levin, Perrot et al. 2000; Reynolds 2000), in labora-
tory animals (Bjrkman, Hughes et al. 1998; Nagaev, Bjrkman et al. 2001),
and in humans (Sherman, Mdluli et al. 1996; Bjrkholm, Sjlund et al.
2001).
The mechanism behind compensation usually involves mutations in the
active site of the target protein that will restore the efficiency of the protein
back to a wild type level. The exact physiological mechanisms by which
compensatory mutations restore fitness have been described in a few cases
(Bjrkman, Samuelsson et al. 1999; Nagaev, Bjrkman et al. 2001). For
example, the cost associated with rpsL mutations causing streptomycin resis-
tance in S. typhimurium, could be compensated by extragenic mutations that
restore the translation efficacy to wild type or nearly wild type levels
(Bjrkman, Samuelsson et al. 1999; Nagaev, Bjrkman et al. 2001).
The degree to which compensation may restore the fitness varies greatly.
Some mutations fully restore the fitness, whereas others only restore the
fitness partially. Bjrkman et al. showed that the selective conditions might
play an important role determining the type of compensatory mutation
achieved. When selecting for compensated rpsL mutants in vivo, only in-
tragenic mutations were detected, whereas selection in vitro only conferred
extragenic suppressor mutations (Bjrkman, Hughes et al. 1998; Bjrkman,
Samuelsson et al. 1999; Nagaev, Bjrkman et al. 2001).
To date, evidence for compensatory evolution occurring in bacteria iso-
lated from humans has been presented for isoniazid-resistant Mycobacterium
23
tuberculosis (Sherman, Mdluli et al. 1996), fusidic acid-resistant Staphylo-
coccus aureus (Nagaev, Bjrkman et al. 2001) and for clarithromycin-
resistant Helicobacter pylori (Bjrkholm, Sjlund et al. 2001). To what ex-
tent compensation occurs in clinical settings and stabilizes resistant patho-
gens is of great medical interest. In the current thesis, we show that clinical
compensation may occur for clarithromycin-resistant Helicobacter pylori.
This was accomplished by isolating a susceptible pre-treatment strain and a
resistant post-treatment strain that were clonally related. Thereafter, the fit-
ness difference of the clinical pair was established and compared with the
fitness divergence of a defined in vitro pair containing the same resistance
mutation and for which the possibilities for compensation had been mini-
mized. By showing that the fitness difference in the clinical pair was smaller
than for the in vitro pair, we conclude that compensatory evolution had re-
duced the fitness cost (Bjrkholm, Sjlund et al. 2001).
An additional way of restoring the fitness is by true reversion of the resis-
tance mutation. This mechanism, which requires a back mutation at the spe-
cific site of the resistance mutation, is not as likely to occur as compensatory
mutations at other sites. This can be attributed to the higher rates of compen-
satory mutations relative to that of reversion, and to the presence of popula-
tion bottlenecks, where the most frequent genotype is more likely to be
transferred instead of the most fit (Levin, Perrot et al. 2000). Moreover, once
a compensated mutant has evolved in a population, it is highly unlikely that
the mutant will revert to a susceptible state. Schrag and colleagues showed
that when a streptomycin-resistant rpsL allele was replaced by a wild type
rpsL
+
allele in evolved fitness-compensated rpsL strains, the resulting strep-
tomycin-susceptible bacteria were less fit than wild type, uncompensated
rpsL mutants or fitness-compensated mutants (Schrag, Perrot et al. 1997). In
other words, genetic compensation establishes an adaptive valley, which
makes it difficult to return to the uncompensated, streptomycin-susceptible
genotype. This phenomenon is illustrated in Figure 2.
24
Figure 2. Stability of antibiotic resistant bacteria. (Modified after Bjrkman J. 2000,
The biological cost of resistance). Antibiotic resistant mutants (AbR) can be as fit as
(i) or less fit (ii) than the susceptible wildtype (AbS). Fitness may be restored by true
reversion (iii) or by compensatory evolution (AbR*)(iv). Compensatory mutations
are more likely to occur than true reversion.
FURTHER FACTORS INFLUENCING THE
STABILITY OF ANTIBIOTIC RESISTANCE
While antibiotics appear to be the main force behind the selection of antibi-
otic resistant bacteria, they are not totally responsible for their persistence
and spread. Once selected, there are several factors that will influence the
stability and maintenance of resistant bacteria in the absence of antibiotics.
One major mechanism has already been mentioned, compensatory evolution.
Other important factors contributing to making resistance development an
irreversible process are the genetic linkage between resistance genes and no-
cost associated types of resistance.
Time
(iii)
(ii)
AbS AbR
AbR
AbR*
Fitness
(i)
(iv)
25
Genetic Linkage of Resistance Genes
Genetic linkage between resistance genes will result in co-selection of the
genes, i.e. the selection and spread of a certain resistance gene due to co-
selection with another resistance gene. Multiple resistance genes are fre-
quently found on plasmids and transposons, and the use of any of the antibi-
otics will result in selection for all the others. For example, despite a signifi-
cant decrease in the use of sulphonamides in the UK from 1991 to 1999, the
frequency of resistance to sulphonamides remained high (1991; 39.7%,
1999; 46.0%), due to linkage of sulphonamide resistance to other resistance
genes, that were under continued selective pressure (Enne, Livermore et al.
2001). Moreover, in poultry, since vanA can be co-selected with erm(B) in
Enterococcus hirae isolates (Borgen, Sorum et al. 2002), despite the exclu-
sion of avoparcin from animal feed, vancomycin resistance can be main-
tained from the use of macrolides. This mechanism may not only obscure the
relationship between antibiotic use and resistance development but will also
act to stabilize resistant populations.
No-cost Associated Resistance Mutations
No-cost associated mutations will leave the resistant bacteria as fit as their
susceptible counterparts and make them just as competitive. However, most
of chromosomally encoded mutations conferring resistance are costly to the
bacterium, although mutants conferring no measurable cost both in vitro and
in vivo have been described. One example of a no-cost associated mutation
is the AAA (Lys) AGA (Arg) mutation in rpsL in S. typhimurium, which
leads to streptomycin resistance.
It has been argued that no-cost mutations preferentially appear during se-
lective conditions in vivo and that resistance mutations in clinical isolates
mainly are of this type (Bttger, Springer et al. 1998; Sander, Springer et al.
2002). Using Mycobacterium smegmatis as a model organism, Sander et al.
showed that under natural conditions (clinical isolates), strong selection
pressure seems to exist for resistance mutations, which impose little or no
fitness burden. They further hypothesize that costly resistance mutations
acquired in vivo might only be found when a cost-neutral resistance does not
exist for a given drug (Sander, Springer et al. 2002).
26
THE ROLE OF THE HUMAN NORMAL
MICROBIOTA IN ANTIBIOTIC RESISTANCE
DEVELOPMENT
Antibiotic treatment of bacterial infections not only exerts a selective pres-
sure against the pathogen to which it is directed, but also affects the indige-
nous microbiota. Consequently, regardless of the indications for their use,
antibiotics will likely select for resistance among the indigenous microflora,
with the likelihood of resistance developing depending on factors previously
addressed in this thesis; the rate and pattern of antibiotic use, the mutation
rate of the bacteria, and the horizontal transfer rate of resistance genes. The
effects of antibiotic treatment on the human microbiota and its role in con-
tributing to resistance development are reviewed below.
The Human Microbiota
The human normal microbiota consists of three major bacterial ecosys-
tems, the gastro-intestinal, the cutaneous and that of the upper respiratory
tract. The bacteria of these ecosystems are essential to human health since
they stimulate the immune response, aid in the digestion of food, help me-
tabolize drugs and act as a barrier against invading pathogens. The gastro-
intestinal ecosystem is the best studied and by far the most populous: the
large intestine comprises 10
14
bacteria in total, or 10
11
bacteria/g of fecal
material (Berg 1996). In the intestine, the quantitatively most important
group of bacteria is the anaerobes (10
11
-10
12
CFU/g), whereas enterobacteria
(including E. coli) and enterococci constitute 0.1 to 1% of the bacterial
population (10
6
-10
8
CFU/g) (Berg 1996; Wold 2000).
During antibiotic treatment all three ecosystems will be affected, but to
different extents. Depending on the spectrum of the agent, the degree of ab-
sorption and route of elimination, the effects on the flora will vary. The po-
tential ecological effects of different antimicrobial agents on the human mi-
crobiota have been described in several studies (Brismar, Edlund et al. 1991;
Stark, Adamsson et al. 1996; Adamsson, Nord et al. 1999; Edlund, Alvan et
al. 2000; Edlund, Beyer et al. 2000; Matute, Schurink et al. 2002) and sum-
marized in an article by Sullivan et al. (Sullivan, Edlund et al. 2001). One
common effect noted after antibiotic treatment was alterations in the balance
between anaerobic and aerobic species.
The consequences of having such an ecological imbalance in the flora
might be variable. First, this may result in overgrowth of already present
microorgansims such as yeasts and Clostridium difficile, which may lead to
diarrhea or colitis. Second, an ecological imbalance may cause a reduction of
27
colonization resistance (Vollaard and Clasener 1994), i.e. the ability of the
microbiota to withstand colonization of invading bacteria such as pathogens.
Another possible consequence of antibiotic treatment is resistance develop-
ment within the normal microbiota (Sullivan, Edlund et al. 2001).
Resistance among the normal microbiota is likely to play a very important
role in resistance development of pathogenic bacteria (Andremont 2003).
The tremendous number of bacteria in the microbiota allows for several dif-
ferent resistant mechanisms to develop, constituting a potential reservoir of
resistance genes that subsequently can be transferred to other species
(Courvalin 1994). For example, tetQ, which confers resistance to tetracy-
cline, and erm-genes, which confer resistance to erythromycin, can exchange
among Bacteroides spp and between Bacteroides and other species of the
human colon (Salyers and Amabile-Cuevas 1997; Shoemaker, Vlamakis et
al. 2001; Andremont 2003). Transfer of resistance genes probably also oc-
curs in other ecosystems. For instance, parts of the mosaic PBP genes of S.
pneumoniae, conferring penicillin resistance, are likely to originate from
viridans streptococci, which tend to be more resistant (Dowson, Coffey et al.
1993; Bryskier 2002). Moreover, it is assumed that the mecA gene that ren-
ders S. aureus resistant to all -lactams originates in coagulase-negative
staphylococci (Wu, de Lencastre et al. 2001). Thus, gene exchange through
horizontal transfer mechanisms may represent the major avenue by which
pathogens acquire resistance.
Additionally, in the current thesis, we present evidence for long-term per-
sistence of resistant enterococci and staphylococci in the normal microbiota,
as a direct consequence of a one-week antibiotic therapy (Sjlund, Wreiber
et al. 2003). We also show that the commensals of patients with high antibi-
otic use are significantly more resistant than corresponding strains isolated
from control patients with no antibiotic treatment (Gustafsson, Sjlund et al.
2003). This further underlines the importance of a careful and prudent use of
antibiotics. After all, the susceptible bacteria in the normal microbiota might
be our best allies in combating further development and spread of antibiotic
resistance and are crucial in preventing colonization by resistant pathogens.
A study by Roos and colleagues in Sweden exemplified the importance of a
susceptible commensal flora and how it can be used to prevent colonization
by pathogenic bacteria. They showed that recolonization with viridans strep-
tococci in children with acute otitis media had an interfering activity against
the otopathogens, and that the rate of re-occurrence of acute otitis media
significantly was reduced in children receiving the viridans streptococci
(Roos, Hakansson et al. 2001).
In summary, a susceptible normal microbiota plays a crucial role in pre-
venting the invasion of pathogens as well as resistant bacteria. In order to
maintain the sensitivity of our microbiota and limit further spread of antibi-
otic resistance, the ecological effects of the administration of antibiotics have
to be considered.
28
Figure 3. Overview of the main factors influencing selection and stability of antibiotic resistance.
Resistant
Population
Low-
fitness
Resistant
Population
Susceptible
Population
mutation
acquired
resistance
No cost-associated resistance
Cost-associated
resistance
Compensatory
evolution
Selection of resistance is dependent on:
Mutation rate and presence of mutator strains
Rate and extent of horizontal transfer (including
the impact of the human microbiota acting as a
reservoir of resistance genes)
Population size
Selective pressure
Stability of resistance is dependent on:
The biological cost of resistance
Compensatory evolution
Genetic linkage with other genes undergoing
selection
Selective pressure
Resistant
Population
The normal
microbiota
29
GENERAL AIM
The present thesis is focused on the selection and stability of antibiotic resis-
tance, and is aimed at studying the impact of antibiotic treatment on patho-
genic bacteria as well as the normal human microbiota. We address this by
studying the mechanisms by which resistance appears, as well as by studying
the biological cost and persistence of resistance.
As a model organism for pathogenic bacteria, we have chosen the peptic
ulcer bacterium, Helicobacter pylori. This bacterium infects more than 50%
of the western worlds population and is associated with both a medically
and economically important infection.
As major model organisms for the human normal microbiota, we have
chosen enterococci and staphylococci, since both species are clinically im-
portant and common causes of nosocomial infections. We also address the
effect of long-term antibiotic treatment on the human normal flora and will
here, in addition to enterococci and staphylococci, include streptococci and
the Gram-negative Escherichia coli.
30
SPECIFIC AIMS
The specific aims of the present investigation were:
I. To assess the frequency of spontaneous mutation in H. pylori
and investigate whether mutator strains can be found among
clinical isolates. An additional aim was to establish the bio-
logical cost of clarithromycin resistance in H. pylori and in-
vestigate whether compensatory evolution may act to reduce
such a fitness cost.
II. To assess the importance of high antibiotic consumption as a
selector for antibiotic resistance and elevated mutation fre-
quencies.
III. To investigate the effect of a commonly used antibiotic treat-
ment on the normal flora of Enterococcus spp. focusing on re-
sistance development.
IV. To study the effect of a commonly used antibiotic treatment
on the normal flora of Staphylococcus spp. and establish the
fitness of resistant S. epidermidis by using an in vivo competi-
tion model.
31
MATERIALS AND METHODS
Bacterial Species Studied in the Present Thesis
Helicobacter pylori
Helicobacter pylori is a Gram-negative, spiral-shaped bacterium that colo-
nizes the human gastric mucosa (Dunn, Cohen et al. 1997). About half of the
worlds population is carrying H. pylori, which makes it one of the most
common bacterial infections in humans. The infection is associated with
gastritis and, in a subset of individuals, also peptic ulceration. H. pylori is
further recognized as a major risk factor for the development of gastric can-
cer and has since 1994 been classified as a type I (definite) carcinogen by the
WHO (IARC 1994).
A triple therapy consisting of two antibiotics and an acid suppressing drug
is required and recommended to clear the infection, and usually leads to
complete elimination of the bacteria and to the healing of ulcers (de Boer
and Tytgat 2000). Commonly used antibiotics are clarithromycin, amoxicil-
lin, metronidazole and tetracycline. Resistance development is the major
cause of treatment failure and resistance to all the above-mentioned antibiot-
ics has been described (Dunn, Cohen et al. 1997; de Boer and Tytgat 2000).
In the first paper of this thesis, H. pylori was chosen as a model organism
for pathogenic bacteria and used for studying the mechanisms by which re-
sistance appears, as well as for addressing the stability of resistance.
Enterococcus spp.
The Gram-positive Enterococcus species are normal inhabitants of the
human colon and constitute 0.1 to 1% of the gastro-intestinal flora. Over the
last years, enterococci have emerged as important bacterial pathogens, caus-
ing nosocomial infections such as endocarditis and bacteraemia (Murray and
Weinstock 1999). Enterococci rapidly develop resistance (Murray 1990).
Soon after the introduction of penicillin, resistant clones of enterococci were
32
found, necessitating the use of other antibiotics to treat infections. Currently,
enterococci show resistance towards chloramphenicol, erythromycin, clin-
damycin, tetracycline, aminoglycosides, penicillin, fluoroquinolones and
vancomycin.
Enterococci were included in two papers of the present thesis (paper II
and III) where the effect of antibiotic treatment on the normal flora of en-
terococci was assessed.
Figure 4. Helicobacter pylori
Figure 5. Enterococcus faecalis
33
Staphylococcus epidermidis
Staphylococcus epidermidis belongs to the group of coagulase-negative
staphylococci (CoNS) and represents one of the most prevalent species of
the human cutaneous microbiota (Kloos and Bannerman 1994). Like entero-
cocci, the CoNS have emerged as major nosocomial pathogens, primarily
causing infections associated with implanted medical devices (Huebner and
Goldmann 1999). Coagulase-negative staphylococci have been reported to
be the third most common causative agent of nosocomial infections and the
most frequent cause of nosocomial bloodstream infections. Infections in
immuno-compromised patients are especially problematic and may be diffi-
cult to treat because of acquired resistance of the bacteria. In recent years, S.
epidermidis has become resistant to many commonly used antibiotics and
may be a reservoir for antibiotic resistance genes in hospitals.
In paper IV, the effect of a commonly used antibiotic treatment on the
normal flora of S. epidermidis was assessed, with focus on resistance devel-
opment. We further used a human in vivo model to establish the fitness and
stability of macrolide-resistant S. epidermidis.
Figure 6. Staphylococcus epidermidis
34
-streptococci
-streptococci belong to the group of viridans streptococci, which is a het-
erogeneous group of -hemolytic and non-hemolytic streptococci. The -
streptococci normally colonize the pharynx and the gastrointestinal tract but
can also give rise to various infections. Most common is dental caries caused
by S. mutans and subacute endocarditis. The majority of strains are highly
susceptible to penicillin, which is the first drug of choice against streptococ-
cal infections (Murray 1998). -streptococci were included in paper II of the
present thesis, where resistance development and mutation frequencies in the
normal microbiota were studied.
Escherichia coli
Escherichia coli belongs to the family Enterobacteriaceae, a large and
medically important group of Gram-negative bacteria. Among the genus
Escherichia, E. coli is the most clinically important bacterium associated
with several diseases, such as meningitis, gastroenteritis, urinary tract infec-
tions and sepsis (Murray 1998).
In paper II, in which we examined how extensive use of antibiotics af-
fected the normal microbiota, E. coli was used as an indicator organism for
the gastro-intestinal flora.
35
Bacterial Strains (I-IV)
Paper (I)
The H. pylori isolates used were:
(i) Clinical isolates collected in a clinical treatment trial performed
by Hultn and colleagues (Hulten, Gibreel et al. 1997). Isolates of
H. pylori were obtained from patients before and 3 months after
undergoing an anti-Helicobacter treatment including the mac-
rolide clarithromycin. From this material, paired susceptible pre-
treatment and resistant post-treatment isolates of the same strain
were obtained.
(ii) Clinical isolates collected in a case-control study of gastric cancer
performed by Enroth et al. (Enroth, Kraaz et al. 2000).
(iii) Reference strain 26695
Paper (II)
Isolates of E. coli, enterococci, -streptococci and coagulase-negative
staphylococci (CoNS) were isolated from nostril, pharynx and feces samples
collected from patients at the Center of Cystic Fibrosis (n=18) and Depart-
ment of Haematology (n=18), University Hospital Uppsala, Uppsala, Swe-
den. Strains from primary health care patients (n=30) with no antibiotic use
one year prior to the study were used as controls.
Paper (III)
Enterococci were isolated from fecal samples obtained from 10 patients par-
ticipating in a prospective cohort study examining the eradication of H. py-
lori in dyspeptic patients. Samples were obtained one day before treatment,
immediately after, one year after, and three years after treatment. From each
study patient and sample, 10 independent colonies of enterococci were iso-
lated and verified.
Paper (IV)
Staphylococcus epidermidis was isolated from nostril samples of patients
participating in a prospective cohort study examining the eradication of H.
pylori in dyspeptic patients. Samples were collected one day before treat-
ment, immediately after, one year after, and four years after treatment. From
each study patient and sample, 10 independent colonies of S. epidermidis
were isolated and verified by Gram staining, catalase, oxidase and manni-
tose-trehalose testing.
36
In the competition assay, reference strain RifR35 was included (Gustafsson,
Cars et al. 2003).
DNA Preparation and Polymerase Chain Reaction (I-
IV)
DNA was extracted from the bacterial strains using Amplicore Respiratory
Preparation Kit (Roche) or Dneasy Tissue kit (Qiagen), both according to the
manufacturers' instructions. Polymerase chain reactions were performed
using standard conditions and the PCR Master kit (Roche). The primers used
are listed in Table 3. All amplified products were separated on a 1.5% aga-
rose gel and visualized with ethidium bromide.
Table 3. Primers used in papers (I-IV).
Primer Sequence Method
AP-PCR
H.pylori
5-CACTCGTCGGGAATGCCCT-3 DNA-
fingerprinting
23SF (p18)
23SR (p21)
5-AGTCGGGACCTAAGGCGAG-3
5-TTCCCGCTTAGATGCTTTCAG-3
Detection of
ClaR mutations
rpoB3
rpoB4
5-GACGTGGTCCATACCTGTAC-3
5-CAATTCATGGACCAAGCTAA-3
Detection of
RifR mutations
(CoNS).
rpoB fw
rpoB rw
5- CCACTTAGGTAACCGTCGTA-3
5- TAATCAATCCAATGTTTGGTC -3
Detection of
RifR mutations
(enterococci,
-streptococci).
rpsL fw
rpsL rw
5- TCACCAGCTTTGAACATTGG-3
5- CCGTATTTAGAACGGCCTTG-3
Detection of
SmR mutations
AP-PCR
D11344
5-AGTGAATTCGCGGTGGATGCCA-3 DNA-
fingerprinting
ermb1
ermb2
5-GAAAAGGTACTCAACCAAATA-3
5-AGTAACGGTACTTAAATTGTTTAC-3
Erm(B) detection
ermc1
ermc2
5GCTAATATTGTTTAAATCGTCAATTCC3
5-GGATCAGGAAAAGGACATTT-3
Erm(C) detection
37
Determination of Antibiotic Susceptibilities (II)
Antibiotic susceptibilities were determined by disk diffusion (Oxoid, Ltd.,
England) according to recommendations by the Swedish Reference Group
for Antibiotics (SRGA) and its subcommittee on methodology (SRGA-M).
Minimal Inhibitory Concentration Determinations
(MIC) (I-IV)
Minimal inhibitory concentrations were determined using the E-test (AB
Biodisk, Solna, Sweden), as recommended by the Swedish Reference Group
for Antibiotics (SRGA).
Mutation Frequency Determination (I, II)
The mutation frequency for H. pylori was determined using rifampicin resis-
tance as a marker. For each strain, 10 l (<10
4
cells) of an overnight culture
was inoculated into 20 independent 1 ml cultures (Brucella broth supple-
mented with 5% FCS) and incubated under microaerophilic conditions at
37C for 36-48 h. Viable count was performed on 3 of the tubes. To deter-
mine the number of resistant mutants, the whole amount of each tube was
spread on GC-agar plates containing rifampicin (15 g/l). The plates were
incubated for 5 days and colonies were counted. The mutation frequency
was calculated from the median number of resistant mutants divided by the
viable count.
In mutation frequency determinations, all the mutants present in a given
population are measured, irrespective of whether the mutation events oc-
curred early or late during the growth of the population. By using the median
of the 20 cultures for the respective strain, the impact of any occasional
jackpot cultures was minimized. In addition, to assure that the 10 l of bacte-
ria used to inoculate the 1 ml sample cultures was free from pre-existing
resistant mutants, 30 l of the pre-culture was plated on a selection plate.
The sample cultures were used to measure the mutation frequency only if no
mutants were present in these 30 l.
In a similar manner, the mutation frequency to clarithromycin resistance
was determined using selective plates containing clarithromycin at a concen-
tration of 10 g/l.
In paper II, the mutation frequency to rifampicin resistance was estimated
for E. coli, -streptococci, CoNS and enterococci. The frequency was calcu-
lated from 10 independent cultures from each of the 3 isolates from each
38
patient/control, resulting in 30 cultures for each species from each patient or
control. The cultures were inoculated with 10
3
bacteria, from a fresh pre-
culture, in 0.4 ml Todd-Hewitt broth (Difco). The cultures were incubated
over-night at 35C giving 10
8
-10
9
cfu/ml. The total number of cells was de-
termined by viable count or optical density measurements. Each culture was
spread on a rifampicin-containing blood agar plate (50 mg/L for E. coli and
enterococci; 0.1mg/L for -streptococci and CoNS). Plates were dried and
incubated for 24 h and then colonies were counted. The mutation frequency
for each of the 3 isolates was estimated by taking the median number of re-
sistant mutants from the 10 cultures divided by the total number of bacterial
cells. Since the three isolates of each species from each patient showed simi-
lar mutation frequencies, we calculated a geometric mean mutation fre-
quency for the 30 cultures.
All pre-cultures were controlled for pre-existing mutants by plating 10
4
bacteria on a selective plate. If pre-existing resistant mutants were found, the
culture was discarded.
The -streptococci were also analyzed for their mutation frequency to
streptomycin resistance. Five independent cultures of one strain from each
patient (n=14) and control (n=17) were included. The remaining strains had
to be excluded because of low bacterial density in the overnight culture or
high initial MIC values to streptomycin. Todd-Hewitt broth (50 ml) was
inoculated with 10
3
bacteria from a fresh broth culture and incubated over
night at 35C, in 5% CO
2
. The bacteria were concentrated by centrifugation
at 1400 g for 15 min, and the pellet was applied to blood agar plates contain-
ing 120 mg/L streptomycin, and colonies were counted after 24 h.
Sequencing (II)
The rpoB gene from 4 independent rifampicin-resistant colonies of entero-
cocci, -streptococci and CoNS was sequenced to verify that the resistant
mutants had mutations in the rpoB gene. Primers for the -subunit of rpoB
were designed using Streptococcus pyogenes (GenBank accession number
AJ295718) and Staphylococcus warneri (GenBank accession number
AF325895) sequences. DNA was prepared by using the Dneasy Tissue kit
(Qiagen). The PCR products were sequenced with ABI prism (Applied Bio-
systems, Warrington, UK) and compared to the original strains.
The rpsL gene from 3 independent streptomycin-resistant colonies of -
streptococci was sequenced to verify that the resistant mutants contained
mutations in the rpsL gene. Primers were constructed from the rpsL gene of
S. pyogenes (Genbank accession number AE006493).
39
Statistical Analysis (II)
The statistical significance of the difference in mutation frequency was de-
termined from the geometrical mean by Mann-Whitney test in Statistical
Analyzing System version 8 and was performed by Kasia Grabowska at the
Swedish Institute for Infectious Disease Control. Further comparison of mu-
tation frequency and total DDD was performed with Microsoft Excel.
Mathematical Modelling (I)
Appendix 1
In paper I, we found a relatively high frequency of hypermutable strains of
H. pylori. The presence of such strains in a bacterial population is predicted
to speed up resistance development caused by chromosomal point mutations.
However, in our material we did not find a correlation between an increased
mutation frequency and resistance development. Therefore, we decided to
address this theoretically by constructing a mathematical model and running
computer simulations in the software program Mathematica in order to de-
termine whether a high mutation frequency is likely to increase resistance
development.
By using a mathematical model describing the dynamics of an H. pylori
population, we calculated the probability that resistant bacteria are present
(and eventually become fixed) in an infected stomach when antibiotic treat-
ment is initiated (appendix 1). With this model, we show that the likelihood
of acquired resistance developing is dependent on the population size, the
mutation rate and the rate by which resistant mutants grow.
In the model, we consider an antibiotic-susceptible population of bacteria
(N) that is kept constant in size by the random removal of individuals at the
same rate as growth occurs. In this population, resistant mutants can arise
with a probability of u, which corresponds to the mutation frequency. Since
there is no antibiotic pressure included in this model, the resistant mutants
that appear are counterselected, i.e. are not as fit as wild types and not fa-
vored by natural selection. The counterselection is modelled by giving the
mutants a reduced growth rate; wild type bacteria always grow with the rate
1 whereas mutants grow by the rate 1+s, where s is the selection coefficient
and always is < 0.
By using these basic conditions, we model the upcoming and growth of
resistant mutants in a population of H. pylori. We show that the probability
of resistant mutants appearing is highly dependent on the mutation rate (u)
and population size (N). In figure 5 in Appendix 1, the probability of fixation
of resistant mutants (P
fix
) is plotted as a function of uN, and the impact of
40
different growth rates of the mutants are shown. A realistic population size
of H. pylori infecting a human stomach ranges from 10
7
-10
9
bacteria, and the
mutation frequency to clarithromycin resistance is approximately 10
-9
. These
numbers yield a uN value of 0.01-1, which in Figure 5 demonstrate that the
mutants are close to fixation. However, if the mutation frequency is subject
to a 100-fold increase, a higher uN value is achieved, which, according to
the plot in Figure 5, leads to a fixation of resistant mutants. Thus, theoretical
modelling suggests that a high mutation frequency will lead to an increased
resistance development by increasing the probability that resistant mutants
are present when treatment is initiated.
Appendix 2
A major finding of paper I was that clarithromycin resistance confers a bio-
logical cost and that this cost can be compensated for. The pre- and post-
treatment pairs of H. pylori isolates that we used in our competition experi-
ments were isolated 3 months apart. Assuming that compensatory mutations
were selected after the resistance mutations, we estimated the succession
time from susceptible resistant compensated bacteria in the stomach.
This is described in Appendix 2 where a mathematical model is used to sup-
port our hypothesis that a step-wise selection from susceptible resistant
compensated is possible during the time frame of 3 months.
In this model, we consider three genotypes; 0 is the susceptible wild type,
1 is the resistant and 2 is the compensated. At time 0, antibiotic is added to a
fully susceptible population (n
0
) and only a small number of bacteria [those
which become resistant (n
1
)] survive the treatment. Since the resistance is
costly to carry in the absence of an antibiotic pressure, these bacteria grow
slowly with a rate k
1
. During this growth, a new, compensating mutation can
occur with the probability of u
2
per replication. This new variant grows with
a higher rate k
2
, which always is >k
1
.
In Appendix 2, we estimate the time it will take for the compensated mu-
tant to appear and take over the population (P
to
=1). We present the outcome
of 8 different simulations, each of which has been assigned a certain set of
parameter values. For instance, we model the impact of different initial den-
sities (n
1
0
) and different growth rates (k
1
) of the resistant bacteria. We also
model the impact of different mutation frequencies to compensation (u
2
). For
most of the parameter values we use, compensation occurs and becomes
fixed within 90 days. Thus, according to this model and the assumptions we
use, the transition from susceptible resistant compensated is possible
within the time frame of 3 months.
41
In vivo Competition Model (I)
The animal experiments were conducted using protocols approved by the
Animal Studies Ethical Committee of Stockholm.
In paper I, the bacterial fitness of clarithromycin-resistant H. pylori strains
was estimated in vivo using a mouse model. Since the human Lewis
b
blood
group antigen (Le
b
) functions as a receptor for the bacterias adhesins and
mediates attachment to the gastric cells, a transgenic mouse model express-
ing Le
b
-containing glycoconjugates was used to establish a chronic, high-
density infection of H. pylori. These mice contain a human 1,3/4 fucosyl-
transferase ORF under the control of nucleotides 596 to +21 of a fatty acid
binding protein gene (Falk, Bry et al. 1995).
Equal amounts of ClaS wild type and ClaR mutant of each clonally re-
lated pair were mixed and orally inoculated in a volume of 50 l into 6-10
week old animals. Four animals/dose/time point were used, and the experi-
ments were performed twice. H. pylori were recovered by homogenization
of half of the stomach. Samples were cultured on GC-agar plates containing
vancomycin, polymyxin B and trimethoprim with and without 1g/ml
clarithromycin.
Pulsed-Field Gel Electrophoresis (PFGE) (IV)
Bacterial cells were harvested by centrifugation from 3 ml overnight cultures
in brain-heart infusion broth and resuspended in 3 ml of Tris-HCl buffer, pH
7.6. The bacterial suspension (150 l) was mixed with 150 l 2% agarose
(Sigma, USA) in Tris-HCl buffer and used for making the gel plugs. The
plugs were incubated at 35C overnight in 4 ml of Lysis buffer 1 (6mM Tris-
HCl, pH 7.6, 1M NaCl, 100mM EDTA, pH 7.5, 0.5% Brij 58, 0.2% deoxy-
cholate, 0.5% sodium lauryl sarcosine (Sarcosyl), 1mg/ml lysozyme (Life
Technology, Invitrogen, USA) and 7g/ml lysostaphin (Sigma, USA), then
incubated overnight at 55C in 4 ml of Lysis 2 buffer (1% sodium lauryl
sarcosine (Sarcosyl), 0.5 M EDTA, pH 9.5, 50g/ml Proteinase K (Boe-
hringer Mannheim, Germany). The plugs were washed three times for at
least 30 minutes at 35C in 4 ml of TE-buffer. A 3 mm slice of each gel plug
was incubated overnight at 25C with SmaI (Life Technology, Invitrogen,
USA) and buffer, then placed in the wells of a 1.0% agarose gel (Ultra pure
agarose, Life Technology, Invitrogen, USA), sealed with 1.0 % agarose and
put in 0.5 x TBE buffer. Electrophoresis [Gene Path Electrophoresis System
(BioRad, USA)] was performed with the following conditions: 5-60 s switch
interval with a voltage gradient of 6V/h at an angle of 120 for 23 h. After
electrophoresis, the gel was stained with ethidium bromide for 30 min then
destained in distilled water for 1 h, and the DNA visualized with a UV light
42
scanner (Gel doc 1000) and the restriction fragment profiles interpreted by
comparison with each other, with reference S. aureus strain (NCTC8325),
and with a lambda phage DNA standard (New England Biolabs).
Competition Assay on Human Skin (IV)
The bacterial fitness of S. epidermidis strains was established using a compe-
tition assay on human skin. This method involves inoculation of bacteria on
the forearms of human volunteers, from which the bacteria then are sampled
using a pad attached to a plastic applicator.
In paper IV, the bacteria were sampled using a pad (5x5 cm, 85% viscose
and 15% polyester), moistened by a buffer (PBS pH 7.4 supplemented with
2% Tween 80 and 0.3% lecithin) and attached to an applicator. The applica-
tor was rubbed back and forth 10 times at the site of bacterial application.
The pad then was removed from the applicator, transferred to a sterile plastic
bag (Seward Laboratories, London, UK) and processed together with 20 ml
buffer (PBS, pH 7.4 supplemented with 2% Tween 80 and 0.3% lecithin) in
a Stomacher Lab Blender 400 (Seward Laboratories, London, UK) for 1
min. To detect the recovered bacteria, the processed buffer was plated in 2
dilutions, 0.5 and 3 ml (day 1) or 5 ml (day 3) on media containing erythro-
mycin (100 g/ml) or rifampicin (50 g/ml). After plates were incubated
overnight at 35C in 5% CO
2
, colonies were counted and a competition ratio
was calculated between each clarithromycin-resistant isolate and the refer-
ence strain. These ratios were further used to calculate the relative fitness of
the two clarithromycin-resistant strains from each patient. To assess the total
number of bacteria recovered from the skin, 0.5 ml of buffer was cultured on
non-selective media (Columbia blood agar plates, Difco).
43
RESULTS AND DISCUSSION
The scientific work of the present thesis has focused on development and
stability of antibiotic resistance, and the aim was to study the impact of anti-
biotic treatment on pathogenic bacteria as well as the normal human micro-
biota. We have addressed this by studying the mechanisms by which resis-
tance appears, as well as by studying the biological cost and persistence of
resistance. The major findings of this thesis are presented and discussed be-
low.
Mutator strains are common among clinical isolates of
Helicobacter pylori (I).
In paper I, we assessed the frequency of spontaneous mutation in the gastric
pathogen H. pylori.
Twenty-nine clinical strains of H. pylori were used for mutation fre-
quency testing. For each strain, 20 independent 1 ml cultures were set up and
used for measuring the formation of spontaneous mutations. This was ac-
complished by using rifampicin resistance as a mutational marker and by
plating the entire culture onto media containing rifampicin at a concentration
of 15g/ml. The number of mutants was counted, and the mutation fre-
quency was obtained by dividing the median number of mutants by the total
number of bacteria in the culture.
Our results show that the mutation frequency was unusually high in H.
pylori; 25% of all isolates displayed a similar or higher frequency than En-
terobacteriaceae defective mismatch repair mutants. The observed frequen-
cies varied, however, among the isolates, ranging from 10
-5
to 10
-8
(median
10
-6
).
To assess whether the high mutation frequency was correlated to resis-
tance development, the MIC values to 6 antibiotics were determined for all
isolates included. However, no correlation between a high mutation fre-
quency and the number of resistance determinants was found in this mate-
rial.
44
Figure 7. Mutation frequency to rifampicin resistance for 29 H. pylori isolates. The
dotted lines indicate the approximate mutation frequencies to rifampicin resistance
for wild type and mutator Enterobacteriaceae. Indicated are strains isolated from
peptic ulcer- (stippled), non-ulcer dyspepsia (open) and gastric cancer patients (ver-
tical lines).
The high number of hypermutable strains present among H. pylori clinical
isolates requires explanation. One possible reason for H. pylori having a
generally higher mutation frequency is that H. pylori seems to lack a func-
tional mismatch repair system. Analysis of the two sequenced H. pylori ge-
nomes (J99 and 26695) revealed that the bacterium lacks some of the most
important genes involved in the mismatch repair system, mutH and mutL.
Furthermore, it has been suggested that mutS in H. pylori has another func-
tion than observed for Enterobacteriaceae (Eisen 1998). To verify this, we
constructed a mutS knock out and tested the mutation frequency. In Entero-
45
bacteriaceae, this would confer a 1000-fold increase in mutation frequency.
However, in the present material, no difference was found between wild type
and knock out, which thereby supports that mutS has another function in H.
pylori.
The selective pressure of the surrounding environment acting on the bac-
teria could also explain the high number of hypermutable isolates within this
material. It has been shown both theoretically and experimentally, that a
changing environment facilitates the upcoming of mutator strains; because of
the high mutation rate, such strains can more easily adapt to changes than a
wild type. In the case of H. pylori colonization, the bacteria are constantly
challenged by changing pH values, the epithelial cell turnover, food diges-
tion and the host immune response. Thus, strains that are able to adapt to
these conditions are likely to be at a selective advantage.
Likewise, antibiotic use and selection of resistant mutants can enrich for
bacteria with higher mutation frequencies (Mao, Lane et al. 1997). However,
as mentioned previously, no clear correlation between the level of resistance
and mutation frequency was found in this material, suggesting that the expo-
sure of the bacteria to a continuously changing environment in the gastric
mucosa might play a more significant role in the enrichment for mutators in
H. pylori.
Antibiotic pressure is a selector for elevated mutation
frequencies (II)
Even though under laboratory conditions (Mao, Lane et al. 1997) and in
animal models (Giraud, Matic et al. 2001) selection for antibiotic resistance
can rapidly enrich for mutators, it remains unclear whether clinical use of
antibiotics is an important selector for strains with an increased mutation
frequency. In paper II, we addressed this by examining whether high antibi-
otic use enriched for commensal bacteria with elevated mutation frequen-
cies.
The rationale behind this study and the reason for choosing commensal
bacteria was to examine a situation where the host environment was unal-
tered, and where no pathogenesis occurred. As previously mentioned, a con-
tinuously changing environment due to pathogenesis is predicted to select
for mutators. Thus, to examine a situation where the environmental condi-
tions are relatively constant and where no infectious pathogenesis occurs, we
chose to study the impact of extensive antibiotic use on commensal bacteria.
The resistance patterns and mutation frequencies were studied in four spe-
cies: E. coli, enterococci, -streptococci and CoNS.
When comparing the patients and controls, a significant difference in mu-
tation frequency was noted for E. coli, enterococci and CoNS. Thus, the
46
patients had commensal bacteria with a 3-, 1.8-, and 1.5-fold higher mutation
frequency as compared with the controls (P = 0.0001, 0.016 and 0.012, re-
spectively). For -streptococci, no significant difference (P = 0.74) in muta-
tion frequency to rifampicin resistance between the controls and patients was
evident. To further assess whether -streptococci from patients had an in-
creased mutation frequency, we also measured the mutation frequency to
streptomycin resistance for this group of bacteria. The mutation frequency
was 4.7 x 10
-10
and 1.5 x 10
-10
for patients and controls, respectively (P =
0.024), suggesting that strains with an increased mutation frequency might
be enriched also in the -streptococci. Why this difference was only seen for
streptomycin resistance and not rifampicin resistance is unclear, but might be
related to which types of base pair substitutions cause resistance to the two
antibiotics.
Figure 8. Mutation frequencies to rifampicin resistance of four species isolated from
patients and controls. Each data point represents the mutation frequency of one bac-
terial species from one patient or control. The overall geometric mean mutation
frequency for each bacterial species for the whole patient and control group is indi
cated by a line.
We further examined whether there was a correlation between the level of
resistance to different antibiotic classes and the mutation frequency. For
aminoglycosides, -lactams, macrolides and trimethoprim
sulfamethoxazole, no correlation was observed. In contrast, when comparing
ciprofloxacin-resistant and -susceptible isolates of E. coli within the patient
1,E-09
1,E-08
1,E-07
1,E-06
1,E-05
m
u
t
a
t
i
o
n

f
r
e
q
u
e
n
c
y

t
o

r
i
f
a
m
p
i
c
i
n

r
e
s
i
s
t
a
n
c
e
patients controls patients controls patients controls patients controls
E. coli enterococci -streptococci CoNS
47
group, the mutation frequencies differed significantly. Thus, ciprofloxacin-
resistant E. coli showed a higher mutation frequency than the susceptible
isolates (P = 0.036). Because of this observation, the MICs of ciprofloxacin
were determined for all strains of E. coli and enterococci from patients and
controls. The MICs of ciprofloxacin for E. coli varied between 0.006 and
>32 mg/L for patients and 0.004 and 0.016 mg/L for controls, and this dif-
ference was highly significant (P = 0.002). Similarly, the corresponding
MICs of ciprofloxacin for enterococci were 0.38-32 mg/L for patients and
0.25-3 mg/L for controls (P = 0.0006). Furthermore, we examined whether
antibiotic exposure (DDDs) was correlated with level of resistance or muta-
tion frequency using linear regression. There was no correlation between
total DDD of antibiotics and mutation frequency, nor between DDD of
fluoroquinolones and ciprofloxacin resistance or mutation frequency. Fi-
nally, within each patient there was no correlation between the mutation
frequencies for the different bacterial species.
Even though the increased mutation frequency in the patient group was
significant, the magnitude of the increase was small. These increases could
be caused by mutations of small effect in any of the processes that affect the
mutation frequency in a bacterium (e.g. DNA polymerase, mismatch repair,
excision repair, etc.). Unexpectedly, at least in the light of previous studies
where frequencies of strong mutators were 1.236%, only one such mutator
(an E. coli from the patient group) was found among the 193 commensal
isolates examined. One explanation for the rarity of strong mutators could be
that for these investigated bacterial species, mutators are impaired for growth
in the commensal flora of hosts, thereby preventing them from increasing to
a high frequency. A second potential explanation is that putative mutator
genes were removed by recombination after resistance was acquired. Finally,
a third possibility, and the one we find most likely, is that most of the antibi-
otics given to these patients are poor selectors of mutators. Thus, it is pre-
dicted that mutators are mainly enriched when the resistance is conferred by
a chromosomal mutation (e.g. gyrA) that is formed at a higher rate in the
mutator than the non-mutator. Most of the resistances in the examined bacte-
ria result from genes that are located on horizontally transferred genetic ele-
ments (e.g. plasmids), and a mutator is not expected to increase the rate of
plasmid transfer. The only exception is fluoroquinolone resistance, which is
caused by sequential chromosomal mutations that confer stepwise increases
in resistance. Thus, acquiring high-level fluoroquinolone resistance is pre-
dicted to enrich for mutators. Support for this idea comes from the finding
that a significant correlation between resistance and increased mutation fre-
quency was only seen for ciprofloxacin.
48
Clarithromycin resistance confers a biological cost in
Helicobacter pylori but can be reduced by
compensatory evolution in vivo (I)
To address whether clarithromycin resistance confers a cost in H. pylori, we
performed competition experiments between isogenic susceptible and
clarithromycin-resistant isolates. The experiment was performed both in
vitro and in vivo. By using a transgenic mouse model, expressing the human
blood group antigen Lewis
b
on the surface of the mucus cells, we were able
to establish a chronic infection in the mouse stomach. Five different clonally
related pairs of clinical isolates were used, each one consisting of a suscepti-
ble pre-treatment isolate and a resistant post-treatment isolate. Further, we
performed competition experiments for one pair for which the resistant iso-
late was isolated in vitro. In each experiment, the mice were inoculated with
a mixture of the bacteria in a 1:1 ratio, and after certain time points, the CFU
was determined by plating on selective and non-selective media. The compe-
tition index (CI) was calculated and the fitness associated with the resistance
mutation estimated.
For two of the clinical pairs, G6-G44 and G47-G108, for unknown reasons
no infection could be established in the mice. Thus, only three of the clinical
pairs and the defined pair isolated in vitro were used. Comparison of the
fitness in mice of the pair ClaS (67:21)-ClaR (67:21, A
2143
=> G) with the
pair ClaS (G142)-ClaR (G193, A
2143
=> G) suggested that the cost of
clarithromycin resistance had been reduced in the G193 mutant. The ratios
of ClaS/ClaR bacteria at day 7 were 1167 and 14, respectively, indicating
that most of the fitness cost of the A2143G mutation had been lost in strain
G193. An even clearer example of the reduction in cost was demonstrated by
the comparison between the ClaS (G34)-ClaR (G49, A
2142
=> G) and ClaS
(G83)-ClaR (G162, A
2142
=> G) pairs. For the G34/G49 pair the post-
treatment resistant mutant lost against the susceptible strain; the competition
ratio was 337 at day 7. In contrast, in the second pair G83/G162, the post-
treatment resistant mutant out-competed the susceptible pre-treatment strain;
the competition ratio was 0.004 at day 7. Thus, in this case the cost of the
A2142G mutations was completely eliminated in strain G162.
At least two explanations can be proposed for the different costs of an
identical resistance mutation in different isolates. One possibility, as has
been shown to occur both in laboratory media and in experimental animals,
is that a compensatory mutation was selected after the resistance mutation
occurred, resulting in a strain with increased fitness. Another possibility is
that the cost depends on the genetic background of the clinical isolates in
which the resistance mutation appeared. However, with respect to the stabi-
lization of resistance mutations, it is less important whether the compensa-
tory mutations were present before or appeared after the resistance mutation,
49
because in both cases the cost would be reduced and result in stabilization of
the resistant clone in the population.
Assuming that compensatory mutations were selected after the resistance
mutation, we estimated the succession time from susceptible resistant
compensated bacteria in the stomach. This was accomplished by using theo-
retical modelling and simulations in silico. From this analysis, we estimated
a take over time of 40-120 days, similar to the 90 days between the isolation
of the pre- and post-treatment strains. Thus, by using the model and the un-
derlying assumptions described in Appendix 2, we present calculations that
are consistent with the possibility that there was a stepwise selection from
susceptible resistant compensated mutant during the time between
sampling of the pre- and post-treatment strains.
Antibiotic treatment selects for stable resistant strains in
the human normal microbiota (II, III, IV)
Antimicrobial agents are known to influence the human microbiota, and the
extent of disturbance depends on a variety of factors, including drug dosage,
route of administration and the pharmacokinetic/dynamic properties of the
agent. Even though the microbiota may return to normal rapidly after com-
pletion of a treatment, this thesis work presents evidence for long-term per-
sistence (several years) of selected resistant commensal bacteria. Such per-
sistence and the natural exchange of genes between bacteria make the human
microbiota a reservoir of resistance genes for potential spread to pathogens.
In paper II, we show that extensive antibiotic use selects for commensals
with highly increased resistance. A clear shift towards more resistance was
seen among isolates from antibiotic-treated patients for virtually all types of
resistances examined. For example, the frequency of ciprofloxacin resistance
in enterococci and CoNS was 63% and 67%, respectively, as compared to
0% in the controls. Similarly, the frequency of erythromycin-resistant CoNS
was 52% in the patient group and 0% in the controls. Resistance to tobramy-
cin was uncommon in E. coli and CoNS, in spite of a rather high use of to-
bramycin. Finally, vancomycin and imipenem resistance was not detected at
all in the patient group, but these drugs were also the least used.
The pronounced shift towards increased resistance among the patient iso-
lates was expected. However, we cannot say whether the resistant clones
were pre-existing in the individuals, selected de novo during treatment or
acquired from the hospital environment, but it is clear that the resistant
strains were enriched during treatment to eventually dominate the micro-
flora.
In paper III, the effect of a commonly used antibiotic treatment used to
eradicate H. pylori on the normal microbiota of enterococci was assessed. In
50
this study, five patients were included, all colonized with H. pylori and hav-
ing either a duodenal or gastric ulcer for which a seven-day course of
clarithromycin 250 mg b.i.d., metronidazole 400 mg b.i.d. and omeprazole
20 mg b.i.d. was indicated. Subjects who previously had been treated to
eradicate H. pylori, or who had received antibiotic treatment within the prior
4 weeks were not included. Five other patients, who also had dyspeptic
symptoms, but who did not receive any antibiotic treatment, were used as
controls. During the study period, no further antibiotic treatment was initi-
ated. In the 10 subjects, fecal samples were collected 1 day before, 3-7 days
immediately after, one year, and three years after treatment. From each study
patient and from each sample, 10 independent colonies were isolated on bile
esculin agar, restreaked on blood agar plates and their DNA fingerprint, MIC
to clarithromycin, and erm(B) presence determined.
Results showed that a one-week treatment of H. pylori, consisting of
clarithromycin, metronidazole, and omeprazole, selects for highly macrolide
resistant enterococci and that these can persist at least 3 years post-treatment
in the absence of any further selection. Three of the five treated patients pos-
sessed highly resistant enterococci one year after the treatment was ended. In
one of these patients, the resistant population persisted for three years.
Similarly, we assessed the effects of the same treatment regimen on the
flora of S. epidermidis (paper IV). Samples from the nares of five patients
were collected one day before, 3-7 days immediately after, and one and four
years after treatment. From each study patient and each sample, 10 inde-
pendent colonies of S. epidermidis were isolated on Columbia blood agar
plates (Difco) and verified by Gram staining, positive catalase, negative
Dnase, negative mannitose and negative trehalose testing.
One day prior to treatment, all five patients harbored clarithromycin-
susceptible (MIC <0.5 g/ml) S. epidermidis among the 10 independent
colonies examined. In four patients, all 10 isolates were susceptible, but in
one patient, two isolates (of 10) were highly resistant due to the presence of
erm(C). Immediately after completing treatment, four of five patients dis-
played high-level clarithromycin-resistant (MIC >256 g/ml) isolates. The
other isolates from this time point were either resistant with lower MIC val-
ues (16-96g/ml) or susceptible. Highly resistant isolates still could be de-
tected one year after treatment in four patients and in three of five patients
four years post-treatment. By PCR, all highly resistant isolates harbored
erm(C).
To explain the long-term persistence of resistance in this study, we inves-
tigated whether macrolide-resistant S. epidermidis changed in fitness over
the 4-year study period. By performing competition experiments between
isogenic clarithromycin-resistant isolates and a susceptible reference strain
on the forearms of human volunteers, we provided evidence that fitness of
resistant S. epidermidis increased. Whether this increase in fitness is due to
genetic compensation of the cost associated with erm(C) carriage cannot be
51
concluded from the current data, but serves as a possible explanation to our
findings. The Erm(C) methylase in S. epidermidis is encoded on a 2.3 kb
plasmid; however, the cost of erm(C) to S. epidermidis is not known.
To summarize, antibiotic treatment affects our indigenous microbiota and
can potentially give rise to long-term colonization with resistant populations.
Long-term persistence of resistant enterococci and staphylococci is impor-
tant for several reasons. First, both S. epidermidis and enterococci belong to
the group of bacteria that occasionally can cause severe disease, even though
they are part of the normal microbiota. Therefore, stably resistant popula-
tions increase risk of unsuccessful treatment of such infections. Second, re-
sistance in the normal microbiota might contribute to increased resistance
development among pathogens by interspecies transfer of resistant determi-
nants. Since the population size of the normal microbiota is so large, it is
possible for multiple different resistant variants to develop, which increase
the risk of spread to populations of pathogenic bacteria. Persistent resistant
populations in the normal microbiota further enhance the risk of transfer,
especially if the selecting antibiotic is used for treatment.
52
CLINICAL IMPLICATIONS
In the current thesis, we present evidence that compensation of the cost of
antibiotic resistance occurs in a clinical setting. Only two former reports
have described such compensation in vivo. The implication of in vivo com-
pensation is that resistant pathogens may be more difficult to get rid of than
initially anticipated. Since compensatory evolution makes the resistant iso-
lates just as fit as their susceptible counterparts, they are likely to persist and
remain for a long time, even in the absence of an antibiotic pressure.
The presence of such stable resistant bacteria may confer problems for indi-
vidual patients as well for the society in general. Treatment of infections
caused by resistant bacteria usually takes longer than infections caused by
susceptible bacteria and is thus more costly. Moreover, resistant infections
have been associated with higher mortality rates in intensive care patients.
In the present thesis, we further show that clinical use of antibiotics se-
lects for stable resistance in the human microflora. This is important for sev-
eral reasons. First, many commensals occasionally can cause severe disease,
even though they are part of the normal microbiota. Therefore, stable resis-
tant populations increase the risk of unsuccessful treatment of such infec-
tions. Second, resistance in the normal microbiota might contribute to in-
creased resistance development among pathogens by interspecies transfer of
resistant determinants.
In conclusion, our results provide insight into how resistance is stabilized
in a bacterial population and show that the normal microbiota may be an
important reservoir of resistance genes. These findings accentuate the need
for strategies to prevent resistance development. Therefore, a restrictive and
more prudent use of antibiotics should continue to be encouraged together
with an improved infection and transmission control. Hopefully this may
serve to reduce resistance and act to prolong the lifespan of our currently
available antibiotics.
53
ACKNOWLEDGEMENTS
During my years as a graduate student I have come across and met a number
of persons to whom I would like to express my sincere gratitude and appre-
ciation:
Lars Engstrand, my advisor, thank you for your never-ending enthusiasm,
for always taking time listening to me and for always supporting me in my
decisions. From you I have learnt the Engstrand-essentials of research,
which can be summarized in three Cs: Cash, Collaboration and Confer-
ences. During my time in your lab, I have, besides science, got to know how
important it is to apply for funding, establish collaborations with other re-
searchers and travel to meetings presenting my data to other scientists. For
this I am truly grateful and I really appreciate all trips to conferences and
meetings that you have made possible. I also wish to thank you for all the
fun times during these trips and for taking your students to the best shopping
malls when abroad.
Dan I. Andersson, my co-advisor, for sharing your extensive knowledge in
bacterial genetics with me, for always coming up with solutions to problems
and for being a fun, fika-loving professor who preferentially meets with his
students at Caf Fgelsngen.
Professor Bruce R. Levin, Dept of Biology, Emory University, Atlanta,
GA, for letting me work in your lab, for introducing me to mathematical
modelling as well as American blues, and for being a real source of inspira-
tion during my years as a graduate student.
Professor Martin J. Blaser, New York University School of Medicine, for
good collaboration and great comments on two manuscripts.
Camilla Wiuff, my Danish friend who I got to know during my stay in At-
lanta, GA. You have been a great support to me, both in science and in life.
Thanks for many fun times in the lab in Atlanta and for great trips to Savan-
nah and Florida. Most of all thanks for always being someone to count on,
and for being a true and honest friend.
Ingegerd Gustafsson, IG, my friend and former colleague, thanks for intro-
ducing me to clinical bacteriology when I first came to the lab in Uppsala
and for being the best research-pal in antibiotic resistance.
54
Svena Svenska McGill, for looking after and taking care of me in Atlanta
and for letting me stay with you in your house on Kay Lane. Thanks for all
fun times in Atlanta; Jelly rolls, pancakes, hash browns, Whitewater, con-
certs, Swedish parties, Braves games, Ben & Jerrys, Lenox Mall, Phipps
Mall, laser shows, Jimmy Buffett, Cadavar balls, BBQs You are the best!
The Bact Lab Babes, THE colleagues of all colleagues, we have had so
much fun together so I do not even know where to begin We have worked
(and sung), travelled (and sung), skied (and sung), danced (and sung) and
maybe most of all laughed together and it has been great fun. My sincere
thanks to Sandra Hjalmarsson, Ingegerd Gustafsson, Maria Held, Caroline
Fock, Sara Olofsson, Helene Kling, Svena McGill, Marie Edvinsson,
Johanna Henriksns, Kristine Marteliusson and Elisabeth Nielsen. Included
in the Bact Lab Babes are also the Bact Lab Boys. A big thanks to Calle
Rubin, Martin Storm, Thomas Cars, Christian Ehrenborg, Patrik Jern and
Amarinder Bindra for being cool guys and fun to be around.
Pylori-Pinglorna, THE pinglor of all pinglor, my dear co-workers and
friends in Stockholm, thanks for being so kind and sweet, Anna Skoglund,
Christina Persson, Christina Nilsson, Marianne Ljungstrm, Annelie Lundin,
Britta Bjrkholm, Helene Kling, Karin Wreiber, Hedvig Engstrm, Mathilda
Lindberg, Kristina Schnmeyr and Lena Eriksson. Also a big thanks to Far-
zad Olfath and Mrten Kivi for being fun and good colleagues.
Past and present members of the Cars Lab, especially Otto Cars; an en-
thusiastic and dedicated researcher who is a great source of inspiration to
graduate students. Thanks for many fun times singing together in Larz-Ottoz
and for letting me come with you and your group on ski-trips.
Astrid Asklin and Eva Tano; two strong ladies from Norrbotten who have
helped me and supported me through the years. Thanks to Eva for good col-
laboration, for teaching me PFGE and for rewarding my hard work in the lab
by giving me the very prestigious title Norrlands Stolthet. My warmest
thanks to Astrid for sharing your extensive knowledge in bacteriology with
me and for teaching me the basis of bacteriological identification and classi-
fication.
All the people at Clinical Bacteriology in Uppsala, especially Anita
Perols, Marius Domeika, Eva Haxton, Bjrn Herrmann, Ulla Zettersten,
Margareta Edvall, Monica Eriksson, Erik Torell, Ulrika Ransj, Eva Hjelm
and Angela Lagerqvisth-Widh.
All the people at the Swedish Institute for Infectious Disease Control,
especially past and present members of Professor Anderssons group, all
members of the TB-groups, especially my friend Emma Huitric.
Britt-Marie Hoffmann for being the best (and definitely the most trendy)
coordinator ever, having an excellent taste for fashion and style as well as a
fantastic mind for social events.
Fellow graduate student Carolina Johansson for being the best roomie
and friend during our stay at Cold Spring Harbor. Patricia Komp-Lindgren
55
for having a blast together in Istanbul, I still cannot stop laughing when I
think about how we ended up at that fancy five star hotel. Sandra Hjal-
marsson, Caroline Fock, Helene Kling and Fredrik Bckhed for fun
times hiking around in Australia. Helena Enroth for fun and nice company
in Orlando and at Rikstmman.
All good friends of mine and my old classmates from Ume and Uppsala.
Special thanks to Linda Lindh and Helene Snigel Kling for being fun
and good friends and to the girls Susanna Lindfors, Lisa Pettersson and
Anna Persson (a.k.a Miranda, Charlotte and Samantha), all with an abso-
lutely brilliant mind for serious shopping and for great fun both in Sweden
and NYC! Thanks to Mikael Sandgren and Nicklas Une for fun times
during Christmas holidays spent in Lycksele. Sofia Dahlberg, many thanks
for friendship and good collaboration.
Finally, I wish to thank my family, my dad Berth, my mum Viveca, my
sister sa and her husband Urban, my nephew Oscar and my grandmother
Ingrid for their encouragement and support.
Last, but not least, my sincere thanks to Jona Karlsson, my biggest sup-
porter and in whom I put all my trust. Thank you kindly for being the won-
derful person you are. I love you always.
Maria
56
REFERENCES
Adamsson, I., C. E. Nord, et al. (1999). "Comparative effects of omeprazole,
amoxycillin plus metronidazole versus omeprazole, clarithromycin
plus metronidazole on the oral, gastric and intestinal microflora in
Helicobacter pylori-infected patients." J Antimicrob Chemother
44(5): 629-40.
Andersson, D. I. (2003). "Persistence of antibiotic resistant bacteria." Curr
Opin Microbiol 6(5): 452-6.
Andersson, D. I. and B. R. Levin (1999). "The biological cost of antibiotic
resistance." Curr Opin Microbiol 2(5): 489-93.
Andremont, A. (2003). Commensal flora may play key role in spreading
antibiotic resistance. ASM News. 69: 601-607.
Arason, V. A., K. G. Kristinsson, et al. (1996). "Do antimicrobials increase
the carriage rate of penicillin resistant pneumococci in children?
Cross sectional prevalence study." Bmj 313(7054): 387-91.
Arjan, J. and G. M. de Visser (2002). "The fate of microbial mutators." Mi-
crobiology 148: 1247-1252.
Asensio, A., A. Guerrero, et al. (1996). "Colonization and infection with
methicillin-resistant Staphylococcus aureus: associated factors and
eradication." Infect Control Hosp Epidemiol 17(1): 20-8.
Austin, D. J., K. G. Kristinsson, et al. (1999). "The relationship between the
volume of antimicrobial consumption in human communities and
the frequency of resistance." Proc Natl Acad Sci U S A 96(3): 1152-
6.
Ball, P., F. Baquero, et al. (2002). "Antibiotic therapy of community respira-
tory tract infections: strategies for optimal outcomes and minimized
resistance emergence." J Antimicrob Chemother 49(1): 31-40.
Baquero, F. (1997). "Gram-positive resistance: challenge for the develop-
ment of new antibiotics." J Antimicrob Chemother 39 Suppl A: 1-6.
Barbosa, T. M. and S. B. Levy (2000). "The impact of antibiotic use on re-
sistance development and persistence." Drug Resist Updat 3(5): 303-
311.
Berg, R. D. (1996). "The indigenous gastrointestinal microflora." Trends
Microbiol 4(11): 430-5.
Bjrkholm, B., M. Sjlund, et al. (2001). "Mutation frequency and biological
cost of antibiotic resistance in Helicobacter pylori." Proc Natl Acad
Sci U S A 98(25): 14607-12.
Bjrkman, J. and D. I. Andersson (2000). "The cost of antibiotic resistance
from a bacterial perspective." Drug Resist Updat 3(4): 237-245.
57
Bjrkman, J., D. Hughes, et al. (1998). "Virulence of antibiotic-resistant
Salmonella typhimurium." Proc Natl Acad Sci U S A 95(7): 3949-
53.
Bjrkman, J., I. Nagaev, et al. (2000). "Effects of environment on compensa-
tory mutations to ameliorate costs of antibiotic resistance." Science
287(5457): 1479-82.
Bjrkman, J., P. Samuelsson, et al. (1999). "Novel ribosomal mutations af-
fecting translational accuracy, antibiotic resistance and virulence of
Salmonella typhimurium." Mol Microbiol 31(1): 53-8.
Blondeau, J. M. (2004). "Fluoroquinolones: mechanism of action, classifica-
tion, and development of resistance." Surv Ophthalmol 49 Suppl 2:
S73-8.
Borgen, K., M. Sorum, et al. (2002). "Genetic linkage between erm(B) and
vanA in Enterococcus hirae of poultry origin." Microb Drug Resist
8(4): 363-8.
Bridges, B. A. (2001). "Hypermutation in bacteria and other cellular sys-
tems." Philos Trans R Soc Lond B Biol Sci 356(1405): 29-39.
Brismar, B., C. Edlund, et al. (1991). "Comparative effects of clarithromycin
and erythromycin on the normal intestinal microflora." Scand J In-
fect Dis 23(5): 635-42.
Bronzwaer, S. L., O. Cars, et al. (2002). "A European study on the relation-
ship between antimicrobial use and antimicrobial resistance." Emerg
Infect Dis 8(3): 278-82.
Bryskier, A. (2002). "Viridans group streptococci: a reservoir of resistant
bacteria in oral cavities." Clin Microbiol Infect 8(2): 65-9.
Burman, L. a. O.-L., B (2001). A Global Perspective on Bacterial Infections,
Antibiotic Usage, and the Antibiotic Resistance Problem. Antibiotic
Development and Resistance. D. a. A. Hughes, D.I. London, Taylor
and Francis Inc.: 1-21.
Bttger, E. C., B. Springer, et al. (1998). "Fitness of antibiotic-resistant mi-
croorganisms and compensatory mutations." Nat Med 4(12): 1343-4.
Cars, O., S. Mlstad, et al. (2001). "Variation in antibiotic use in the Euro-
pean Union." Lancet 357(9271): 1851-3.
Chambers, H. F. (2001). "The changing epidemiology of Staphylococcus
aureus?" Emerg Infect Dis 7(2): 178-82.
Chopra, I., A. J. O'Neill, et al. (2003). "The role of mutators in the emer-
gence of antibiotic-resistant bacteria." Drug Resist Updat 6(3): 137-
45.
Clark, N. M., E. Hershberger, et al. (2003). "Antimicrobial resistance among
gram-positive organisms in the intensive care unit." Curr Opin Crit
Care 9(5): 403-12.
Cohen, M. L. (1992). "Epidemiology of drug resistance: implications for a
post-antimicrobial era." Science 257(5073): 1050-5.
Cohen, M. L. (2000). "Changing patterns of infectious disease." Nature
406(6797): 762-7.
58
Courvalin, P. (1994). "Transfer of antibiotic resistance genes between gram-
positive and gram-negative bacteria." Antimicrob Agents Chemother
38(7): 1447-51.
Davies, J. (1994). "Inactivation of antibiotics and the dissemination of resis-
tance genes." Science 264(5157): 375-82.
de Boer, W. A. and G. N. Tytgat (2000). "Regular review: treatment of
Helicobacter pylori infection." Bmj 320(7226): 31-4.
Dowson, C. G., T. J. Coffey, et al. (1993). "Evolution of penicillin resistance
in Streptococcus pneumoniae; the role of Streptococcus mitis in the
formation of a low affinity PBP2B in S. pneumoniae." Mol Micro-
biol 9(3): 635-43.
Dunn, B. E., H. Cohen, et al. (1997). "Helicobacter pylori." Clinical Micro-
biology Reviews 10(4): 720-741.
Edlund, C., G. Alvan, et al. (2000). "Pharmacokinetics and comparative ef-
fects of telithromycin (HMR 3647) and clarithromycin on the oro-
pharyngeal and intestinal microflora." J Antimicrob Chemother
46(5): 741-9.
Edlund, C., G. Beyer, et al. (2000). "Comparative effects of moxifloxacin
and clarithromycin on the normal intestinal microflora." Scand J In-
fect Dis 32(1): 81-5.
Eisen, J. A. (1998). "A phylogenomic study of the MutS family of proteins."
Nucleic Acids Res 26(18): 4291-300.
Enne, V. I., D. M. Livermore, et al. (2001). "Persistence of sulphonamide
resistance in Escherichia coli in the UK despite national prescribing
restriction." Lancet 357(9265): 1325-8.
Enroth, H., W. Kraaz, et al. (2000). "Helicobacter pylori strain types and risk
of gastric cancer: a case-control study." Cancer Epidemiol Bio-
markers Prev 9(9): 981-5.
Falk, P. G., L. Bry, et al. (1995). "Expression of a human alpha-1,3/4-
fucosyltransferase in the pit cell lineage of FVB/N mouse stomach
results in production of Leb-containing glycoconjugates: a potential
transgenic mouse model for studying Helicobacter pylori infection."
Proc Natl Acad Sci U S A 92(5): 1515-9.
Finch, R. G. (1998). "Antibiotic resistance." J Antimicrob Chemother 42(2):
125-8.
Giraud, A., I. Matic, et al. (2002). "Mutator bacteria as a risk factor in treat-
ment of infectious diseases." Antimicrob Agents Chemother 46(3):
863-5.
Giraud, A., I. Matic, et al. (2001). "Costs and benefits of high mutation rates:
adaptive evolution of bacteria in the mouse gut." Science 291(5513):
2606-8.
Giraud, A., M. Radman, et al. (2001). "The rise and fall of mutator bacteria."
Curr Opin Microbiol 4(5): 582-5.
Goettsch, W., W. van Pelt, et al. (2000). "Increasing resistance to fluoroqui-
nolones in escherichia coli from urinary tract infections in the neth-
erlands." J Antimicrob Chemother 46(2): 223-8.
59
Granizo, J. J., L. Aguilar, et al. (2000). "Streptococcus pyogenes resistance
to erythromycin in relation to macrolide consumption in Spain
(1986-1997)." J Antimicrob Chemother 46(6): 959-64.
Green, D. W. (2002). "The bacterial cell wall as a source of antibacterial
targets." Expert Opin Ther Targets 6(1): 1-19.
Gustafsson, I., O. Cars, et al. (2003). "Fitness of antibiotic resistant Staphy-
lococcus epidermidis assessed by competition on the skin of human
volunteers." J Antimicrob Chemother 52(2): 258-63.
Gustafsson, I., M. Sjlund, et al. (2003). "Bacteria with increased mutation
frequency and antibiotic resistance are enriched in the commensal
flora of patients with high antibiotic usage." J Antimicrob Chemo-
ther 52(4): 645-50.
Hakenbeck, R. (1999). "beta-lactam-resistant Streptococcus pneumoniae:
epidemiology and evolutionary mechanism." Chemotherapy 45(2):
83-94.
Hand, W. L. (2000). "Current challenges in antibiotic resistance." Adolesc
Med 11(2): 427-38.
Hawkey, P. M. (2000). "Mechanisms of resistance to antibiotics." Intensive
Care Med 26 Suppl 1: S9-13.
Hooper, D. C. (1999). "Mechanisms of fluoroquinolone resistance." Drug
Resist Updat 2(1): 38-55.
Hooper, D. C. (2001). "Minimizing potential resistance: the molecular view-
-a comment on Courvalin and Trieu-Cuot." Clin Infect Dis 33 Suppl
3: S157-60.
Huebner, J. and D. A. Goldmann (1999). "Coagulase-negative staphylo-
cocci: role as pathogens." Annu Rev Med 50: 223-36.
Hulten, K., A. Gibreel, et al. (1997). "Macrolide resistance in Helicobacter
pylori: mechanism and stability in strains from clarithromycin-
treated patients." Antimicrob Agents Chemother 41(11): 2550-3.
IARC (1994). "Infection with Helicobacter pylori." IARC Monogr Eval Car-
cinog Risks Hum 61: 177-240.
Jones, R. N. (2001). "Resistance patterns among nosocomial pathogens:
trends over the past few years." Chest 119(2 Suppl): 397S-404S.
Kloos, W. E. and T. L. Bannerman (1994). "Update on clinical significance
of coagulase-negative staphylococci." Clin Microbiol Rev 7(1): 117-
40.
Komp Lindgren, P., A. Karlsson, et al. (2003). "Mutation rate and evolution
of fluoroquinolone resistance in Escherichia coli isolates from pa-
tients with urinary tract infections." Antimicrob Agents Chemother
47(10): 3222-32.
LeClerc, J. E., B. Li, et al. (1996). "High mutation frequencies among Es-
cherichia coli and Salmonella pathogens." Science 274(5290): 1208-
11.
Levin, B. R., V. Perrot, et al. (2000). "Compensatory mutations, antibiotic
resistance and the population genetics of adaptive evolution in bacte-
ria." Genetics 154(3): 985-97.
60
Levy, S. B. (2001). "Antibiotic resistance: consequences of inaction." Clin
Infect Dis 33 Suppl 3: S124-9.
Li, Z., C. Kelley, et al. (1998). "Expression of katG in Mycobacterium tu-
berculosis is associated with its growth and persistence in mice and
guinea pigs." J Infect Dis 177(4): 1030-5.
Lieberman, J. M. (2003). "Appropriate antibiotic use and why it is impor-
tant: the challenges of bacterial resistance." Pediatr Infect Dis J
22(12): 1143-51.
Low, D. E. (2001). "Antimicrobial drug use and resistance among respira-
tory pathogens in the community." Clin Infect Dis 33 Suppl 3:
S206-13.
Maiden, M. C. (1998). "Horizontal genetic exchange, evolution, and spread
of antibiotic resistance in bacteria." Clin Infect Dis 27 Suppl 1: S12-
20.
Mao, E. F., L. Lane, et al. (1997). "Proliferation of mutators in A cell popu-
lation." J Bacteriol 179(2): 417-22.
Martinez, J. L. and F. Baquero (2000). "Mutation frequencies and antibiotic
resistance." Antimicrob Agents Chemother 44(7): 1771-7.
Matute, A. J., C. A. Schurink, et al. (2002). "Double-blind, placebo-
controlled study comparing the effect of azithromycin with
clarithromycin on oropharyngeal and bowel microflora in volun-
teers." Eur J Clin Microbiol Infect Dis 21(6): 427-31.
Miller, J. H. (1996). "Spontaneous mutators in bacteria: insights into path-
ways of mutagenesis and repair." Annu Rev Microbiol 50: 625-43.
Miller, K., A. J. O'Neill, et al. (2002). "Response of Escherichia coli hyper-
mutators to selection pressure with antimicrobial agents from differ-
ent classes." J Antimicrob Chemother 49(6): 925-34.
Miller, K., A. J. O'Neill, et al. (2004). "Escherichia coli mutators present an
enhanced risk for emergence of antibiotic resistance during urinary
tract infections." Antimicrob Agents Chemother 48(1): 23-9.
Moellering, R. C. (2003). "Linezolid: the first oxazolidinone antimicrobial."
Ann Intern Med 138(2): 135-42.
Murray, B. E. (1990). "The life and times of the Enterococcus." Clin Micro-
biol Rev 3(1): 46-65.
Murray, B. E. and G. M. Weinstock (1999). "Enterococci: new aspects of an
old organism." Proc Assoc Am Physicians 111(4): 328-34.
Murray, P. R., Rosenthal, K.S., Kobayashi, G.S. and Pfaller, M.A. (1998).
Enterobacteriaceae. Medical Microbiology. M. Brown. St. Louis,
Mosby, Inc.: 232-244.
Murray, P. R., Rosenthal, K.S., Kobayashi, G.S. and Pfaller, M.A. (1998).
Streptococcus. Medical Microbiology. M. Brown. St. Louis, Mosby,
Inc.: 189-205.
Nagaev, I., J. Bjrkman, et al. (2001). "Biological cost and compensatory
evolution in fusidic acid-resistant Staphylococcus aureus." Mol Mi-
crobiol 40(2): 433-9.
Normark, B. H. and S. Normark (2002). "Evolution and spread of antibiotic
resistance." J Intern Med 252(2): 91-106.
61
Norrby, R. (2001). "Linezolid--a review of the first oxazolidinone." Expert
Opin Pharmacother 2(2): 293-302.
Norrby, R. a. C., O (2003). Antibiotika- och Kemoterapi. Stockholm, Liber
AB.
Oliver, A., R. Canton, et al. (2000). "High frequency of hypermutable Pseu-
domonas aeruginosa in cystic fibrosis lung infection." Science
288(5469): 1251-4.
Orencia, M. C., J. S. Yoon, et al. (2001). "Predicting the emergence of anti-
biotic resistance by directed evolution and structural analysis." Nat
Struct Biol 8(3): 238-42.
Pankey, G. A. and L. D. Sabath (2004). "Clinical relevance of bacteriostatic
versus bactericidal mechanisms of action in the treatment of Gram-
positive bacterial infections." Clin Infect Dis 38(6): 864-70.
Pittet, D. (2003). "Hand hygiene: improved standards and practice for hospi-
tal care." Curr Opin Infect Dis 16(4): 327-35.
Ren, L., M. S. Rahman, et al. (1999). "Escherichia coli cells exposed to
streptomycin display a mutator phenotype." J Bacteriol 181(3):
1043-4.
Reynolds, M. G. (2000). "Compensatory evolution in rifampin-resistant Es-
cherichia coli." Genetics 156(4): 1471-81.
Rice, L. B. (2000). "Bacterial monopolists: the bundling and dissemination
of antimicrobial resistance genes in gram-positive bacteria." Clin In-
fect Dis 31(3): 762-9.
Roos, K., E. G. Hakansson, et al. (2001). "Effect of recolonisation with "in-
terfering" alpha streptococci on recurrences of acute and secretory
otitis media in children: randomised placebo controlled trial." Bmj
322(7280): 210-2.
Salyers, A. A. and C. F. Amabile-Cuevas (1997). "Why are antibiotic resis-
tance genes so resistant to elimination?" Antimicrob Agents Chemo-
ther 41(11): 2321-5.
Sander, P., B. Springer, et al. (2002). "Fitness cost of chromosomal drug
resistance-conferring mutations." Antimicrob Agents Chemother
46(5): 1204-11.
Schaaff, F., A. Reipert, et al. (2002). "An elevated mutation frequency favors
development of vancomycin resistance in Staphylococcus aureus."
Antimicrob Agents Chemother 46(11): 3540-8.
Schmid, M. B. (2001). New Targets and Strategies for Identification of
Novel Classes of Antibiotics. Antibiotic Development and Resis-
tance. D. a. A. Hughes, D.I. London, Taylor and Francis Inc.: 197-
209.
Schrag, S. J., V. Perrot, et al. (1997). "Adaptation to the fitness costs of anti-
biotic resistance in Escherichia coli." Proc R Soc Lond B Biol Sci
264(1386): 1287-91.
Seppl, H., T. Klaukka, et al. (1997). "The effect of changes in the con-
sumption of macrolide antibiotics on erythromycin resistance in
group A streptococci in Finland. Finnish Study Group for Antim-
icrobial Resistance." N Engl J Med 337(7): 441-6.
62
Sherman, D. R., K. Mdluli, et al. (1996). "Compensatory ahpC gene expres-
sion in isoniazid-resistant Mycobacterium tuberculosis." Science
272(5268): 1641-3.
Shoemaker, N. B., H. Vlamakis, et al. (2001). "Evidence for extensive resis-
tance gene transfer among Bacteroides spp. and among Bacteroides
and other genera in the human colon." Appl Environ Microbiol
67(2): 561-8.
Sjlund, M., K. Wreiber, et al. (2003). "Long-term persistence of resistant
Enterococcus species after antibiotics to eradicate Helicobacter py-
lori." Ann Intern Med 139(6): 483-7.
Sniegowski, P. D., P. J. Gerrish, et al. (1997). "Evolution of high mutation
rates in experimental populations of E. coli." Nature 387(6634): 703-
5.
Stark, C. A., I. Adamsson, et al. (1996). "Effects of omeprazole and amoxy-
cillin on the human oral and gastrointestinal microflora in patients
with Helicobacter pylori infection." J Antimicrob Chemother 38(6):
927-39.
Stratton, C. W. (2003). "Dead bugs don't mutate: susceptibility issues in the
emergence of bacterial resistance." Emerg Infect Dis 9(1): 10-6.
Sullivan, A., C. Edlund, et al. (2001). "Effect of antimicrobial agents on the
ecological balance of human microflora." Lancet Infect Dis 1(2):
101-14.
Taddei, F., I. Matic, et al. (1997). "To be a mutator, or how pathogenic and
commensal bacteria can evolve rapidly." Trends Microbiol 5(11):
427-8; discussion 428-9.
Taddei, F., M. Radman, et al. (1997). "Role of mutator alleles in adaptive
evolution." Nature 387(6634): 700-2.
Tanabe, K., T. Kondo, et al. (1999). "A conspicuous adaptability to antibiot-
ics in the Escherichia coli mutator strain, dnaQ49." FEMS Microbiol
Lett 176(1): 191-6.
Tenaillon, O., B. Toupance, et al. (1999). "Mutators, population size, adap-
tive landscape and the adaptation of asexual populations of bacte-
ria." Genetics 152(2): 485-93.
Tenover, F. C. and J. M. Hughes (1996). "The challenges of emerging infec-
tious diseases. Development and spread of multiply-resistant bacte-
rial pathogens." Jama 275(4): 300-4.
Williams, J. D. (2001). "Antibiotic resistance in hospital pathogens--
acquisition or spread?" Int J Antimicrob Agents 18(3): 295-8.
Wold, A. E. a. A., I (2000). Pathological Consequences of Commensalism.
Persistent Bacterial Infections. J. Nataro, Blaser, MJ and Cunning-
ham-Rundles, S. Washington DC, ASM Press: 115-144.
Vollaard, E. J. and H. A. Clasener (1994). "Colonization resistance." Antim-
icrob Agents Chemother 38(3): 409-14.
Wu, S. W., H. de Lencastre, et al. (2001). "Recruitment of the mecA gene
homologue of Staphylococcus sciuri into a resistance determinant
and expression of the resistant phenotype in Staphylococcus aureus."
J Bacteriol 183(8): 2417-24.
63
Ysern, P., B. Clerch, et al. (1990). "Induction of SOS genes in Escherichia
coli and mutagenesis in Salmonella typhimurium by fluoroqui-
nolones." Mutagenesis 5(1): 63-6.
Acta Universitatis Upsaliensis
Comprehensive Summaries of Uppsala Dissertations
from the Faculty of Medicine
Editor: The Dean of the Faculty of Medicine
Distribution:
Uppsala University Library
Box 510, SE-751 20 Uppsala, Sweden
www.uu.se, acta@ub.uu.se
ISSN 0282-7476
ISBN 91-554-6026-7
A doctoral dissertation from the Faculty of Medicine, Uppsala University,
is usually a summary of a number of papers. A few copies of the complete
dissertation are kept at major Swedish research libraries, while the sum-
mary alone is distributed internationally through the series Comprehen-
sive Summaries of Uppsala Dissertations from the Faculty of Medicine.
(Prior to October, 1985, the series was published under the title Abstracts of
Uppsala Dissertations from the Faculty of Medicine.)